bump pointer: internal, .claude/commands (TPM-PSS distress, D-370)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

.claude/CLAUDE.md

@@ -38,6 +38,8 @@
 - Simplicity first: minimal, targeted changes. No temporary fixes — find the root cause.
 - Where architecture is flawed, state is duplicated, or patterns are inconsistent, propose and implement the structural fix. Ask what a perfectionist reviewer would reject, and fix that too.
 - On non-trivial changes, ask whether there is a more elegant way before presenting. Skip it for simple fixes — don't over-engineer.
+- Reuse before you write: search for an existing helper by what it does (its key call, literal or regex), not by a guessed name. Web helpers live in `lib/` (route-only ones in a `lib/` beside the route), backend helpers in `Functions/Helpers/`.
+- Dead and copied code is gated: the .NET build fails on unused private members and parameters (`.editorconfig`), `npm run knip` covers web and MCP, and `npm run dup` in `tools/quality` fails on new clones. CI only reports, so run them before calling a change done.
 - English for code comments and everything that carries knowledge.
 
 Private half of these instructions (process, working files, technical docs, operations): @../internal/CLAUDE.md

.editorconfig

@@ -0,0 +1,11 @@
+# Analyzer severities for the .NET build. Directory.Build.props runs the IDE rules in
+# `dotnet build` (EnforceCodeStyleInBuild), and the warning freeze makes every warning an error.
+root = true
+
+[*.cs]
+# Unused private members, private members that are written but never read, and unused parameters
+# of non-public methods. An injected dependency nobody reads is a feature that silently never runs.
+dotnet_diagnostic.IDE0051.severity = warning
+dotnet_diagnostic.IDE0052.severity = warning
+dotnet_diagnostic.IDE0060.severity = warning
+dotnet_code_quality_unused_parameters = non_public

.github/dependabot.yml

@@ -65,7 +65,7 @@ updates:
       - dependency-name: '*'
         update-types: ['version-update:semver-major']
 
-  # Validator tooling for rules/ (ajv-cli); combine-rules.yml installs it from this
+  # Validator tooling for rules/ (ajv for validate.js); combine-rules.yml installs it from this
   # lockfile with contents:write in hand, so the pins must not go stale.
   - package-ecosystem: npm
     directory: /rules/scripts
@@ -81,6 +81,23 @@ updates:
         applies-to: security-updates
         patterns: ['*']
 
+  # Duplicate-check tooling (jscpd for tools/quality/check-duplicates.mjs), installed by the
+  # CI duplicates job from this lockfile. A jscpd update can change the clone fingerprints:
+  # if its PR fails the job, regenerate the baseline (npm run dup:baseline) in that PR.
+  - package-ecosystem: npm
+    directory: /tools/quality
+    schedule:
+      interval: weekly
+    cooldown:
+      default-days: 7
+    open-pull-requests-limit: 5
+    groups:
+      quality-tooling:
+        patterns: ['*']
+      quality-tooling-security:
+        applies-to: security-updates
+        patterns: ['*']
+
   # Every action is referenced by commit SHA with the version in a trailing comment;
   # action-pins.yml fails any workflow change that reintroduces a mutable tag. Dependabot
   # is what keeps those pins current: it rewrites the SHA and the comment together, so the

.github/scripts/Assert-HostsFromRegistry.ps1

@@ -0,0 +1,84 @@
+<#
+.SYNOPSIS
+    Fails when a workflow or a CI helper spells out a host instead of reading Get-Endpoints.ps1.
+
+.DESCRIPTION
+    The CI counterpart of the URL-registry guard tests (HardcodedUrlGuardTests,
+    hardcodedUrls.guard.test.ts, hardcoded-urls.guard.test.ts). Enforced: every host the registry
+    defines -- derived from it, so the list cannot fall behind -- plus any Azure storage, App Service
+    or autopilotmonitor.com host, so a new or misspelled account is caught as well.
+
+    Scans .github/workflows/*.yml and .github/scripts/*.ps1 except the registry itself. Comment lines
+    and PowerShell block comments do not count: comments may cite URLs.
+
+.PARAMETER Root
+    Repository root to scan. Defaults to the parent of .github/scripts.
+#>
+[CmdletBinding()]
+param(
+    [string]$Root
+)
+
+$ErrorActionPreference = 'Stop'
+
+if (-not $Root) {
+    $Root = (Resolve-Path (Join-Path $PSScriptRoot '../..')).Path
+}
+
+$endpoints = & (Join-Path $PSScriptRoot 'Get-Endpoints.ps1')
+$registryHosts = $endpoints.Values | ForEach-Object { ([uri]$_).Host } | Sort-Object -Unique
+$patterns = @($registryHosts | ForEach-Object { [regex]::Escape($_) }) + @(
+    '[a-z0-9-]+\.(blob|table|queue|file)\.core\.windows\.net'
+    '[a-z0-9-]+\.azurewebsites\.net'
+    '([a-z0-9-]+\.)*autopilotmonitor\.com'
+)
+$hostPattern = [regex]::new('(' + ($patterns -join '|') + ')', 'IgnoreCase')
+
+$files = @(
+    Get-ChildItem -Path (Join-Path $Root '.github/workflows') -File -Filter '*.yml'
+    Get-ChildItem -Path (Join-Path $Root '.github/workflows') -File -Filter '*.yaml'
+    Get-ChildItem -Path (Join-Path $Root '.github/scripts') -File -Filter '*.ps1' |
+        Where-Object Name -ne 'Get-Endpoints.ps1'
+)
+
+$findings = 0
+foreach ($file in $files) {
+    $relative = [System.IO.Path]::GetRelativePath($Root, $file.FullName) -replace '\\', '/'
+    $isPowerShell = $file.Extension -eq '.ps1'
+    $inBlock = $false
+    $lineNumber = 0
+    foreach ($line in [System.IO.File]::ReadAllLines($file.FullName)) {
+        $lineNumber++
+        $code = $line
+        if ($isPowerShell) {
+            if ($inBlock) {
+                $end = $code.IndexOf('#>')
+                if ($end -lt 0) { continue }
+                $code = $code.Substring($end + 2)
+                $inBlock = $false
+            }
+            $start = $code.IndexOf('<#')
+            if ($start -ge 0) {
+                $end = $code.IndexOf('#>', $start + 2)
+                if ($end -lt 0) {
+                    $code = $code.Substring(0, $start)
+                    $inBlock = $true
+                } else {
+                    $code = $code.Remove($start, $end + 2 - $start)
+                }
+            }
+        }
+        if ($code.TrimStart().StartsWith('#')) { continue }
+
+        $match = $hostPattern.Match($code)
+        if ($match.Success) {
+            Write-Host "::error file=$relative,line=$lineNumber::$($match.Value) is spelled out here -- take it from .github/scripts/Get-Endpoints.ps1."
+            $findings++
+        }
+    }
+}
+
+if ($findings -gt 0) {
+    throw "$findings host literal(s) outside the endpoint registry."
+}
+Write-Host "No host literal outside .github/scripts/Get-Endpoints.ps1 ($($files.Count) files scanned)."

.github/scripts/Get-Endpoints.ps1

@@ -0,0 +1,71 @@
+<#
+.SYNOPSIS
+    The one place the workflows and the release scripts take their hosts from.
+
+.DESCRIPTION
+    CI's member of the URL registry (Shared Constants.cs, web lib/config.ts, MCP src/config.ts):
+    every own host the workflows, the .github/scripts helpers and the local release build talk to,
+    plus the Authenticode timestamp server, is defined here once. Assert-HostsFromRegistry.ps1
+    (Workflow gates) fails on any of these hosts spelled out elsewhere under .github/.
+
+    The storage URLs derive from the account name, so blob and table can never point at two
+    different accounts.
+
+    A file in the repository rather than a GitHub variable: the local release build reads it too,
+    and bearer tokens are sent to these URLs -- where a token goes stays a reviewed line in git.
+
+.PARAMETER ExportToGitHubEnv
+    Write every entry to $GITHUB_ENV instead of returning it, so the following steps of the job
+    read them as environment variables ($env:AGENT_CONTAINER_URL, ${{ env.AGENT_CONTAINER_URL }}).
+
+.OUTPUTS
+    [ordered] hashtable, keyed by the names the environment variables carry.
+
+.EXAMPLE
+    $endpoints = & "$PSScriptRoot/Get-Endpoints.ps1"
+    Invoke-WebRequest -Uri "$($endpoints.AGENT_CONTAINER_URL)/version.json"
+#>
+[CmdletBinding()]
+param(
+    [switch]$ExportToGitHubEnv
+)
+
+$ErrorActionPreference = 'Stop'
+
+$storageAccount = 'autopilotmonitoreu'
+# Pre-migration account in another tenant; only the fail-soft legacy mirror still writes there.
+$legacyStorageAccount = 'autopilotmonitor'
+
+$blob  = "https://$storageAccount.blob.core.windows.net"
+$table = "https://$storageAccount.table.core.windows.net"
+
+$endpoints = [ordered]@{
+    STORAGE_BLOB_URL           = $blob
+    STORAGE_TABLE_URL          = $table
+    AGENT_CONTAINER_URL        = "$blob/agent"
+    VERSIONS_CONTAINER_URL     = "$blob/versions"
+    STATS_CONTAINER_URL        = "$blob/stats"
+    # Platform settings row; CI writes only its agent and bootstrap version columns.
+    ADMIN_CONFIG_URL           = "$table/AdminConfiguration(PartitionKey='GlobalConfig',RowKey='config')"
+    LEGACY_AGENT_CONTAINER_URL = "https://$legacyStorageAccount.blob.core.windows.net/agent"
+    # Front Door in front of the agent container: what customer scripts download from.
+    DOWNLOAD_ALIAS_URL         = 'https://download.autopilotmonitor.com/agent'
+    API_BASE_URL               = 'https://autopilotmonitor-api-eu.azurewebsites.net'
+    MCP_BASE_URL               = 'https://mcp.autopilotmonitor.com'
+    PORTAL_BASE_URL            = 'https://portal.autopilotmonitor.com'
+    WWW_BASE_URL               = 'https://www.autopilotmonitor.com'
+    DOCS_BASE_URL              = 'https://docs.autopilotmonitor.com'
+    CODE_SIGN_TIMESTAMP_URL    = 'http://timestamp.acs.microsoft.com'
+}
+
+if ($ExportToGitHubEnv) {
+    if ([string]::IsNullOrWhiteSpace($env:GITHUB_ENV)) {
+        throw '-ExportToGitHubEnv only works inside a GitHub Actions step.'
+    }
+    $endpoints.GetEnumerator() | ForEach-Object { "$($_.Key)=$($_.Value)" } |
+        Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+    $endpoints.GetEnumerator() | ForEach-Object { Write-Host "$($_.Key) = $($_.Value)" }
+    return
+}
+
+return $endpoints

.github/scripts/Publish-BootstrapScripts.ps1

@@ -81,7 +81,6 @@ param(
     [string]$SigningToken,
     [string]$KeyVaultUrl,
     [string]$CertificateName,
-    [string]$TimestampUrl = 'http://timestamp.acs.microsoft.com',
     [string]$ExpectedPublisher,
     [string]$SignedOutputDir,
     [switch]$DryRun
@@ -93,10 +92,13 @@ if (-not $RepoRoot) {
     $RepoRoot = (Resolve-Path (Join-Path $PSScriptRoot '../..')).Path
 }
 
-$ContainerUrl       = 'https://autopilotmonitoreu.blob.core.windows.net/agent'
-$LegacyContainerUrl = 'https://autopilotmonitor.blob.core.windows.net/agent'
-$AliasUrl           = 'https://download.autopilotmonitor.com/agent'
-$TableUrl           = "https://autopilotmonitoreu.table.core.windows.net/AdminConfiguration(PartitionKey='GlobalConfig',RowKey='config')"
+# Every host, the timestamp server included, from the one endpoint registry.
+$Endpoints          = & (Join-Path $PSScriptRoot 'Get-Endpoints.ps1')
+$ContainerUrl       = $Endpoints.AGENT_CONTAINER_URL
+$LegacyContainerUrl = $Endpoints.LEGACY_AGENT_CONTAINER_URL
+$AliasUrl           = $Endpoints.DOWNLOAD_ALIAS_URL
+$TableUrl           = $Endpoints.ADMIN_CONFIG_URL
+$TimestampUrl       = $Endpoints.CODE_SIGN_TIMESTAMP_URL
 
 $BootstrapSource   = Join-Path $RepoRoot 'scripts/Bootstrap/Install-AutopilotMonitor.ps1'
 $BootstrapBlob     = 'Install-AutopilotMonitor.ps1'
@@ -210,7 +212,7 @@ function Set-PublishSetSignature {
         --timestamp-rfc3161 $TimestampUrl `
         --file-digest sha256 `
         --description 'Autopilot Monitor bootstrap' `
-        --description-url 'https://www.autopilotmonitor.com' `
+        --description-url $Endpoints.WWW_BASE_URL `
         $paths
 
     # Native tool: PowerShell does not throw on a non-zero exit, and an unnoticed signing
@@ -327,8 +329,10 @@ $loaderVersion = Get-BootstrapScriptVersion -Content $loaderContent -Origin $Loa
 Write-Host "Loader script version: $loaderVersion"
 
 # ------------------------------------------------------------------ 2. dev render
-$stableUrlLiteral = '$AgentDownloadUrl = "https://download.autopilotmonitor.com/agent/AutopilotMonitor-Agent.zip"'
-$devUrlLiteral    = '$AgentDownloadUrl = "https://download.autopilotmonitor.com/agent/AutopilotMonitor-Agent-dev.zip"'
+# The anchors are the customer scripts' own text, built from the alias in the registry. Should
+# the alias ever move, the scripts have to move with it -- until they do, the render stops here.
+$stableUrlLiteral = '$AgentDownloadUrl = "' + $AliasUrl + '/AutopilotMonitor-Agent.zip"'
+$devUrlLiteral    = '$AgentDownloadUrl = "' + $AliasUrl + '/AutopilotMonitor-Agent-dev.zip"'
 $stableManLiteral = '$VersionJsonName = "version.json"'
 $devManLiteral    = '$VersionJsonName = "version-dev.json"'
 
@@ -339,8 +343,8 @@ $devContent = $sourceContent.Replace($stableUrlLiteral, $devUrlLiteral).Replace(
 Write-Host "Rendered $DevBlob (dev agent URL + version-dev.json)"
 
 # The dev loader must fetch the dev bootstrap, or a dev device would run the stable chain.
-$stableLoaderLiteral = '$BootstrapUrl = "https://download.autopilotmonitor.com/agent/Install-AutopilotMonitor.ps1"'
-$devLoaderLiteral    = '$BootstrapUrl = "https://download.autopilotmonitor.com/agent/Install-AutopilotMonitor-Dev.ps1"'
+$stableLoaderLiteral = '$BootstrapUrl = "' + $AliasUrl + '/Install-AutopilotMonitor.ps1"'
+$devLoaderLiteral    = '$BootstrapUrl = "' + $AliasUrl + '/Install-AutopilotMonitor-Dev.ps1"'
 
 if ($loaderContent.IndexOf($stableLoaderLiteral) -lt 0) {
     throw "Loader dev-render: anchor literal missing in $LoaderSource"

.github/workflows/action-pins.yml

@@ -1,12 +1,15 @@
 name: Workflow gates
 
-# Two checks over the workflow files themselves:
+# Three checks over the workflow files and their helper scripts:
 #   * verify-pins -- every third-party action is referenced by a full commit SHA plus a
 #                    version comment (stricter than zizmor's unpinned-uses audit).
 #   * zizmor      -- static analysis of every workflow (permissions, credential
 #                    persistence, template injection, ...). Configuration and the
 #                    reasoned ignores live in zizmor.yml at the repository root; the
 #                    same run is expected locally before a workflow change is committed.
+#   * hosts       -- no workflow or .github/scripts helper spells out a host; every one
+#                    comes from .github/scripts/Get-Endpoints.ps1, CI's member of the URL
+#                    registry (the C#, web and MCP registries have guard tests of their own).
 #
 # Every third-party action must be referenced by a full commit SHA, never by a tag.
 # Tags are mutable: whoever controls an action repository can repoint v4 at new code,
@@ -26,15 +29,18 @@ name: Workflow gates
 #
 # The pins are not maintained by hand — .github/dependabot.yml opens a monthly grouped
 # PR that rewrites both the SHA and the trailing version comment. The comment is
-# therefore required too: it is what makes a pinned line reviewable.
+# therefore required too: it is what makes a pinned line reviewable. It names the exact
+# release tag (# v7.0.1), never the floating major (# v7): a major tag moves with every
+# upstream release, after which the comment no longer names the pinned commit and
+# zizmor's ref-version-mismatch flags every line that carries it.
 
 on:
   push:
     branches: [main]
-    paths: ['.github/workflows/**', '.github/dependabot.yml', 'zizmor.yml']
+    paths: ['.github/workflows/**', '.github/scripts/**', '.github/dependabot.yml', 'zizmor.yml']
   pull_request:
     branches: [main]
-    paths: ['.github/workflows/**', '.github/dependabot.yml', 'zizmor.yml']
+    paths: ['.github/workflows/**', '.github/scripts/**', '.github/dependabot.yml', 'zizmor.yml']
   workflow_dispatch:
 
 permissions:
@@ -45,7 +51,7 @@ jobs:
     name: Verify action pins
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
@@ -59,13 +65,15 @@ jobs:
             # nothing to pin.
             case "$ref" in ./*|\$/*) continue ;; esac
 
-            if [[ "$ref" =~ ^[^@]+@[0-9a-f]{40}[[:space:]]+#[[:space:]]*[^[:space:]] ]]; then
+            if [[ "$ref" =~ ^[^@]+@[0-9a-f]{40}[[:space:]]+#[[:space:]]*v?[0-9]+\.[0-9]+\.[0-9]+([[:space:]]|$) ]]; then
               printf 'ok  %s:%s  %s\n' "$file" "$line" "$ref"
               continue
             fi
 
             if [[ "$ref" =~ ^[^@]+@[0-9a-f]{40}[[:space:]]*$ ]]; then
-              echo "::error file=$file,line=$line::Pinned but the version comment is missing: '$ref'. Write 'owner/action@<sha> # vN' so the line stays reviewable."
+              echo "::error file=$file,line=$line::Pinned but the version comment is missing: '$ref'. Write 'owner/action@<sha> # vX.Y.Z' so the line stays reviewable."
+            elif [[ "$ref" =~ ^[^@]+@[0-9a-f]{40}[[:space:]]+# ]]; then
+              echo "::error file=$file,line=$line::The version comment must name the exact release tag the SHA resolves to ('# vX.Y.Z'), not a floating major: '$ref'. Find it with: gh api repos/<owner>/<action>/git/matching-refs/tags/vN"
             else
               echo "::error file=$file,line=$line::Action '$ref' is not pinned to a commit SHA. Tags are mutable -- resolve it with: gh api repos/<owner>/<action>/git/ref/tags/<tag>"
             fi
@@ -89,7 +97,7 @@ jobs:
       security-events: write   # SARIF upload: findings land in the Code scanning tab next to CodeQL
       actions: read
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
@@ -101,3 +109,17 @@ jobs:
         with:
           version: 1.30.1
           min-severity: low
+
+  hosts:
+    name: Hosts only from the endpoint registry
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
+
+      # The enforced host list is derived from the registry itself, plus every Azure storage,
+      # App Service and autopilotmonitor.com host, so it cannot fall behind a new entry.
+      - name: No host literal outside .github/scripts/Get-Endpoints.ps1
+        shell: pwsh
+        run: ./.github/scripts/Assert-HostsFromRegistry.ps1

.github/workflows/bootstrap-script-gates.yml

@@ -18,7 +18,7 @@ jobs:
     name: Bootstrap guard tests (PS 5.1)
     runs-on: windows-latest
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 

.github/workflows/build-agent.yml

@@ -7,6 +7,14 @@ on:
         description: 'Build number override (leave empty to auto-increment from buildcounter.txt)'
         required: false
         type: string
+      channel:
+        description: 'release = versioned line (+ stable on cutover). dev = lab channel only (Agent-v2-dev.zip, version-v2-dev.json, Agent-dev.zip, version-dev.json); needs publish_as_stable=false and create_release=false'
+        required: true
+        type: choice
+        options:
+          - release
+          - dev
+        default: release
       upload_blob:
         description: 'Upload to versioned namespace (Agent-v2.zip + version-v2.json)'
         required: true
@@ -32,10 +40,11 @@ on:
         required: false
         type: string
 
+# Every job states its own permissions. This default is what the build job runs with: it
+# executes third-party code (MSBuild targets, analyzers, test frameworks) and must never hold
+# more than a read-only token.
 permissions:
-  contents: write
-  id-token: write      # keyless Sigstore signing (artifact attestation)
-  attestations: write  # store the provenance attestation
+  contents: read
 
 # Shared with publish-scripts.yml: both write version.json (this workflow the agent fields,
 # the script publisher bootstrapVersion under If-Match). Serialising them keeps a script push
@@ -48,7 +57,9 @@ env:
   LINE_MAJOR: 2
   AGENT_PROJECT: src/Agent/AutopilotMonitor.Agent.V2/AutopilotMonitor.Agent.V2.csproj
   DIALOG_PROJECT: src/Agent/AutopilotMonitor.SummaryDialog/AutopilotMonitor.SummaryDialog.csproj
-  TEST_PROJECT: src/Agent/AutopilotMonitor.DecisionCore.Tests/AutopilotMonitor.DecisionCore.Tests.csproj
+  DECISIONCORE_TEST_PROJECT: src/Agent/AutopilotMonitor.DecisionCore.Tests/AutopilotMonitor.DecisionCore.Tests.csproj
+  CORE_TEST_PROJECT: src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/AutopilotMonitor.Agent.V2.Core.Tests.csproj
+  DIALOG_TEST_PROJECT: src/Agent/AutopilotMonitor.SummaryDialog.Tests/AutopilotMonitor.SummaryDialog.Tests.csproj
   CSPROJ_PATH: src/Agent/AutopilotMonitor.Agent.V2/AutopilotMonitor.Agent.V2.csproj
   BOOTSTRAP_SOURCE: scripts/Bootstrap/Install-AutopilotMonitor.ps1
   AGENT_OUTPUT: src/Agent/AutopilotMonitor.Agent.V2/bin/Release/net48
@@ -57,25 +68,49 @@ env:
   VERSIONED_ZIP: AutopilotMonitor-Agent-v2.zip
   STABLE_ZIP: AutopilotMonitor-Agent.zip
   STABLE_MANIFEST: version.json
-  # Authenticode via Azure Key Vault; identity and vault come from repository variables.
+  # Lab channel (channel=dev): the names the dev renders of the bootstrap scripts read.
+  CHANNEL: ${{ inputs.channel }}
+  DEV_VERSIONED_ZIP: AutopilotMonitor-Agent-v2-dev.zip
+  DEV_MANIFEST: version-v2-dev.json
+  DEV_STABLE_ZIP: AutopilotMonitor-Agent-dev.zip
+  DEV_STABLE_MANIFEST: version-dev.json
+  # Signed catalog over every other file of the payload. Must equal
+  # Constants.AgentPayloadCatalogFileName; CodeSigningParityTests keeps them in step.
+  PAYLOAD_CATALOG: AutopilotMonitor.Agent.cat
+  # Authenticode via Azure Key Vault; identity and vault come from repository variables, the
+  # timestamp server (CODE_SIGN_TIMESTAMP_URL) from the endpoint registry like every host here.
   # CODE_SIGN_EXPECTED_SUBJECT is optional and pins the publisher in the verify step.
   CODE_SIGN_CLIENT_ID: ${{ vars.CODE_SIGN_CLIENT_ID }}
   CODE_SIGN_CERTIFICATE_NAME: ${{ vars.CODE_SIGN_CERTIFICATE_NAME }}
   CODE_SIGN_KEY_VAULT_URL: ${{ vars.CODE_SIGN_KEY_VAULT_URL }}
   CODE_SIGN_EXPECTED_SUBJECT: ${{ vars.CODE_SIGN_EXPECTED_SUBJECT }}
-  CODE_SIGN_TIMESTAMP_URL: http://timestamp.acs.microsoft.com
 
 jobs:
-  build:
-    runs-on: windows-latest
-    name: Build Agent
+  # Runs only this repository's scripts. The deploy identity trusts the `production`
+  # environment, which admits main only.
+  reserve:
+    name: Reserve build number
+    runs-on: ubuntu-latest
+    environment: production
+    permissions:
+      contents: read
+      id-token: write   # OIDC federated login for the counter write
+    outputs:
+      build_number: ${{ steps.version.outputs.build_number }}
+      version: ${{ steps.version.outputs.version }}
+      tag: ${{ steps.version.outputs.tag }}
 
     steps:
       - name: Checkout repository
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
+      # Every host in this workflow comes from the one endpoint registry.
+      - name: Load endpoints
+        shell: pwsh
+        run: ./.github/scripts/Get-Endpoints.ps1 -ExportToGitHubEnv
+
       # Every released build is signed; a missing variable stops the run before a build
       # number is burnt.
       - name: Check code-signing configuration
@@ -88,33 +123,22 @@ jobs:
           }
           Write-Host "Code-signing configuration present (certificate '$env:CODE_SIGN_CERTIFICATE_NAME')"
 
-      # Logs in FIRST and only to fetch the Key Vault token: azure/login sets the default az
-      # account, so the storage login below has to be the last one. Both federated credentials
-      # are bound to the ref, so this job must never be given an `environment:` -- that would
-      # change the OIDC subject claim and break both logins.
-      - name: Login to Azure (code-signing identity)
-        uses: azure/login@a641126d1b8aa4d1fa005f4f92df94a3a4c4c906 # v3.1.0
-        with:
-          client-id: ${{ env.CODE_SIGN_CLIENT_ID }}
-          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
-          allow-no-subscriptions: true
-
-      - name: Acquire Key Vault token for signing
+      # The dev channel is a lab channel: it never touches stable, AdminConfiguration, the
+      # bootstrap scripts or a GitHub Release. Refused here, before a build number is burnt.
+      - name: Check channel inputs
         shell: pwsh
+        env:
+          PUBLISH_AS_STABLE: ${{ inputs.publish_as_stable }}
+          CREATE_RELEASE: ${{ inputs.create_release }}
         run: |
-          $ErrorActionPreference = 'Stop'
-          $token = az account get-access-token --scope https://vault.azure.net/.default --query accessToken --output tsv
-          if ($LASTEXITCODE -ne 0 -or [string]::IsNullOrWhiteSpace($token)) {
-            throw "Could not acquire a Key Vault token for the code-signing identity."
+          if ($env:CHANNEL -notin @('release', 'dev')) {
+            throw "Unknown channel '$($env:CHANNEL)'."
           }
-          Write-Host "::add-mask::$token"
-          "CODE_SIGN_AKV_ACCESS_TOKEN=$token" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+          if ($env:CHANNEL -eq 'dev' -and ($env:PUBLISH_AS_STABLE -eq 'true' -or $env:CREATE_RELEASE -eq 'true')) {
+            throw "channel=dev publishes the lab channel only: run it with publish_as_stable=false and create_release=false."
+          }
+          Write-Host "Channel: $env:CHANNEL"
 
-      # Every storage write below (counter blob, agent ZIP, manifests, AdminConfiguration)
-      # is authorized by this federated identity -- no SAS secret. One login covers the whole
-      # job: the federated credential is bound to the ref, and a release run takes ~2 minutes,
-      # well inside the token lifetime. Only the legacy mirror still needs a SAS: that account
-      # lives in the pre-migration tenant, where this principal has no identity.
       - name: Login to Azure (OIDC)
         uses: azure/login@a641126d1b8aa4d1fa005f4f92df94a3a4c4c906 # v3.1.0
         with:
@@ -122,18 +146,6 @@ jobs:
           tenant-id: ${{ secrets.AZURE_TENANT_ID }}
           subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}
 
-      # SDK 10 = single build toolchain (parity with local build.ps1 / global.json).
-      - name: Setup .NET
-        uses: actions/setup-dotnet@a98b56852c35b8e3190ac28c8c2271da59106c68 # v6
-        with:
-          dotnet-version: '10.0.x'
-
-      # Version pinned; roll-forward because the tool targets an older runtime than the SDK above.
-      - name: Install AzureSignTool
-        env:
-          DOTNET_ROLL_FORWARD: LatestMajor
-        run: dotnet tool install --no-cache --global AzureSignTool --version 7.0.1
-
       # Build numbers come from the shared counter blob -- one counter and one
       # implementation for CI and the local build (.github/scripts/Request-BuildNumber.ps1).
       # Reserve-before-build with ETag-CAS: concurrent builds can never mint the same
@@ -155,7 +167,7 @@ jobs:
             $override = [int]$env:BUILD_NUMBER_OVERRIDE
           }
 
-          $counterUrl = "https://autopilotmonitoreu.blob.core.windows.net/agent/buildcounter-v$($env:LINE_MAJOR).txt"
+          $counterUrl = "$env:AGENT_CONTAINER_URL/buildcounter-v$($env:LINE_MAJOR).txt"
           $buildNumber = & ./.github/scripts/Request-BuildNumber.ps1 -CounterUrl $counterUrl -Override $override
 
           $csproj = Get-Content $env:CSPROJ_PATH -Raw
@@ -172,22 +184,167 @@ jobs:
           "version=$version" >> $env:GITHUB_OUTPUT
           "tag=v$version" >> $env:GITHUB_OUTPUT
 
+  # Build zone: no login, no secret, no id-token. Its only output is the artifact; the
+  # release job treats that as data and never builds.
+  build:
+    name: Build and test
+    needs: reserve
+    runs-on: windows-latest
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
+
+      # SDK 10 = single build toolchain (parity with local build.ps1 / global.json).
+      - name: Setup .NET
+        uses: actions/setup-dotnet@a98b56852c35b8e3190ac28c8c2271da59106c68 # v6.0.0
+        with:
+          dotnet-version: '10.0.x'
+
       - name: Restore dependencies
         run: |
           dotnet restore ${{ env.AGENT_PROJECT }}
           dotnet restore ${{ env.DIALOG_PROJECT }}
-          dotnet restore ${{ env.TEST_PROJECT }}
+          dotnet restore ${{ env.DECISIONCORE_TEST_PROJECT }}
+          dotnet restore ${{ env.CORE_TEST_PROJECT }}
+          dotnet restore ${{ env.DIALOG_TEST_PROJECT }}
+
+      # Every agent suite runs before the release build. ORDER AND BUILD_NUMBER ARE LOAD-BEARING:
+      # Agent.V2.Core.Tests references the agent exe, so the test build compiles it. With the
+      # reserved number it stamps the release version and the Build Agent step is up to date; run
+      # after that step or without the number, it would rewrite the binaries with another version.
+      - name: Run agent tests (all suites)
+        shell: pwsh
+        env:
+          BUILD_NUMBER: ${{ needs.reserve.outputs.build_number }}
+        run: |
+          foreach ($project in @($env:DECISIONCORE_TEST_PROJECT, $env:CORE_TEST_PROJECT, $env:DIALOG_TEST_PROJECT)) {
+            dotnet test $project --configuration Release --no-restore --verbosity quiet
+            if ($LASTEXITCODE -ne 0) { throw "Tests failed: $project" }
+          }
 
       - name: Build Agent
         run: dotnet build ${{ env.AGENT_PROJECT }} --configuration Release --no-restore
         env:
-          BUILD_NUMBER: ${{ steps.version.outputs.build_number }}
+          BUILD_NUMBER: ${{ needs.reserve.outputs.build_number }}
 
       - name: Build SummaryDialog
         run: dotnet build ${{ env.DIALOG_PROJECT }} --configuration Release --no-restore
 
-      - name: Run Agent unit tests
-        run: dotnet test ${{ env.TEST_PROJECT }} --configuration Release --no-restore --verbosity quiet
+      # One archive per output folder, unsigned. ZipFile rather than Compress-Archive, which
+      # skips hidden files.
+      - name: Package build output
+        shell: pwsh
+        run: |
+          $ErrorActionPreference = 'Stop'
+          Add-Type -AssemblyName System.IO.Compression.FileSystem
+          $outDir = Join-Path $env:RUNNER_TEMP 'agent-unsigned'
+          New-Item -Path $outDir -ItemType Directory -Force | Out-Null
+          [System.IO.Compression.ZipFile]::CreateFromDirectory((Join-Path $env:GITHUB_WORKSPACE $env:AGENT_OUTPUT), (Join-Path $outDir 'agent.zip'))
+          [System.IO.Compression.ZipFile]::CreateFromDirectory((Join-Path $env:GITHUB_WORKSPACE $env:DIALOG_OUTPUT), (Join-Path $outDir 'dialog.zip'))
+          Get-ChildItem $outDir | ForEach-Object { Write-Host "$($_.Name): $([math]::Round($_.Length / 1MB, 2)) MB" }
+
+      - name: Upload unsigned build output
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
+        with:
+          name: agent-unsigned
+          path: ${{ runner.temp }}/agent-unsigned
+          if-no-files-found: error
+          retention-days: 1
+
+  # Trust zone: only this repository's scripts, pinned actions and the pinned signing tool
+  # run here -- never a restore or build. The build output arrives as data. The runner
+  # image's Windows SDK signtool is used read-only, to verify third-party signatures.
+  #
+  # No `environment:` yet: the code-signing credential is still bound to the ref, and a job
+  # carries exactly one OIDC subject, so an environment here would break the signing login.
+  release:
+    name: Sign and publish
+    needs: [reserve, build]
+    runs-on: windows-latest
+    permissions:
+      contents: write      # tag + GitHub Release
+      id-token: write      # code-signing and storage logins, keyless Sigstore signing
+      attestations: write  # store the provenance attestation
+    env:
+      AGENT_VERSION: ${{ needs.reserve.outputs.version }}
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
+
+      # Storage, legacy mirror, docs link, timestamp server: all from the endpoint registry.
+      - name: Load endpoints
+        shell: pwsh
+        run: ./.github/scripts/Get-Endpoints.ps1 -ExportToGitHubEnv
+
+      - name: Download unsigned build output
+        uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
+        with:
+          name: agent-unsigned
+          path: ${{ runner.temp }}/agent-unsigned
+
+      # Only the two expected archives, each into its own fresh output folder; the .NET
+      # extractor refuses entries that would land outside it. Never unpack into the checkout
+      # root: an archive could otherwise replace the scripts this job runs below.
+      - name: Unpack build output
+        shell: pwsh
+        run: |
+          $ErrorActionPreference = 'Stop'
+          Add-Type -AssemblyName System.IO.Compression.FileSystem
+          foreach ($pair in @(@('agent.zip', $env:AGENT_OUTPUT), @('dialog.zip', $env:DIALOG_OUTPUT))) {
+            $archive = Join-Path $env:RUNNER_TEMP "agent-unsigned/$($pair[0])"
+            $target = Join-Path $env:GITHUB_WORKSPACE $pair[1]
+            if (Test-Path $target) { throw "$($pair[1]) already exists -- refusing to unpack over it." }
+            [System.IO.Compression.ZipFile]::ExtractToDirectory($archive, $target)
+            Write-Host "Unpacked $($pair[0]) into $($pair[1])"
+          }
+
+      # Logs in FIRST and only to fetch the Key Vault token: azure/login sets the default az
+      # account, so the storage login below has to be the last one.
+      - name: Login to Azure (code-signing identity)
+        uses: azure/login@a641126d1b8aa4d1fa005f4f92df94a3a4c4c906 # v3.1.0
+        with:
+          client-id: ${{ env.CODE_SIGN_CLIENT_ID }}
+          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
+          allow-no-subscriptions: true
+
+      - name: Acquire Key Vault token for signing
+        shell: pwsh
+        run: |
+          $ErrorActionPreference = 'Stop'
+          $token = az account get-access-token --scope https://vault.azure.net/.default --query accessToken --output tsv
+          if ($LASTEXITCODE -ne 0 -or [string]::IsNullOrWhiteSpace($token)) {
+            throw "Could not acquire a Key Vault token for the code-signing identity."
+          }
+          Write-Host "::add-mask::$token"
+          "CODE_SIGN_AKV_ACCESS_TOKEN=$token" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+
+      # Every storage write below (agent ZIP, manifests, AdminConfiguration) is authorized by
+      # this federated identity -- no SAS secret. Only the legacy mirror still needs a SAS:
+      # that account lives in the pre-migration tenant, where this principal has no identity.
+      - name: Login to Azure (OIDC)
+        uses: azure/login@a641126d1b8aa4d1fa005f4f92df94a3a4c4c906 # v3.1.0
+        with:
+          client-id: ${{ secrets.AZURE_CLIENT_ID }}
+          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
+          subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}
+
+      # Runtime for AzureSignTool only; nothing is built in this job.
+      - name: Setup .NET
+        uses: actions/setup-dotnet@a98b56852c35b8e3190ac28c8c2271da59106c68 # v6.0.0
+        with:
+          dotnet-version: '10.0.x'
+
+      # Version pinned; roll-forward because the tool targets an older runtime than the SDK above.
+      - name: Install AzureSignTool
+        env:
+          DOTNET_ROLL_FORWARD: LatestMajor
+        run: dotnet tool install --no-cache --global AzureSignTool --version 7.0.1
 
       # Authenticode over everything we author; third-party assemblies keep their own signature.
       #
@@ -223,18 +380,17 @@ jobs:
             --timestamp-rfc3161 "$env:CODE_SIGN_TIMESTAMP_URL" `
             --file-digest sha256 `
             --description "Autopilot Monitor Agent" `
-            --description-url "https://www.autopilotmonitor.com" `
+            --description-url "$env:WWW_BASE_URL" `
             ($targets | ForEach-Object { $_.FullName })
 
           if ($LASTEXITCODE -ne 0) {
             throw "azuresigntool failed with exit code $LASTEXITCODE -- aborting before anything is published."
           }
 
-      - name: Create release artifact
-        id: artifact
+      - name: Stage release payload
         shell: pwsh
         run: |
-          $version = "${{ steps.version.outputs.version }}"
+          $version = $env:AGENT_VERSION
           $stagingDir = "staging/agent"
 
           if (-not (Test-Path "$env:AGENT_OUTPUT/AutopilotMonitor.Agent.exe")) {
@@ -276,6 +432,98 @@ jobs:
           }
           Write-Host "Version check passed: EXE = manifest = reserved = $version"
 
+      # Third-party code enters the payload only with its vendor's own signature: every PE that
+      # is not ours needs a valid EMBEDDED Authenticode signature from an allowed vendor. Embedded
+      # only (signtool verify /pa, no /a): Get-AuthenticodeSignature prefers the runner's Windows
+      # catalogs and reports "Microsoft Windows" for files Windows happens to ship as well. An
+      # unsigned file or a new vendor stops the release until a person decides. This guards the
+      # bytes we did not write; a build that alters our own code is D-346's remaining risk.
+      - name: Check third-party signatures
+        shell: pwsh
+        run: |
+          $ErrorActionPreference = 'Stop'
+          $allowedVendors = @('Microsoft Corporation', 'Json.NET (.NET Foundation)')
+
+          $signtool = Get-ChildItem "${env:ProgramFiles(x86)}\Windows Kits\10\bin\*\x64\signtool.exe" -ErrorAction SilentlyContinue |
+            Where-Object { $_.Directory.Parent.Name -match '^\d+(\.\d+){1,3}$' } |
+            Sort-Object { [version]$_.Directory.Parent.Name } -Descending |
+            Select-Object -First 1
+          if (-not $signtool) { throw "signtool.exe not found in the runner image's Windows SDK." }
+          Write-Host "Using $($signtool.FullName)"
+
+          $failed = $false
+          $checked = 0
+          foreach ($file in Get-ChildItem -Path 'staging/agent' -Recurse -File) {
+            # A PE by content (MZ header), not by extension.
+            $head = [byte[]]::new(2)
+            $stream = [System.IO.File]::OpenRead($file.FullName)
+            try { $read = $stream.Read($head, 0, 2) } finally { $stream.Dispose() }
+            if ($read -lt 2 -or $head[0] -ne 0x4D -or $head[1] -ne 0x5A) { continue }
+            if ($file.Name -like 'AutopilotMonitor.*') { continue }   # ours: the verify step pins them
+
+            $checked++
+            $output = & $signtool.FullName verify /pa $file.FullName 2>&1
+            if ($LASTEXITCODE -ne 0) {
+              Write-Host "::error file=$($file.Name)::No valid embedded vendor signature. $(($output | Out-String).Trim())"
+              $failed = $true
+              continue
+            }
+            $subject = [System.Security.Cryptography.X509Certificates.X509Certificate]::CreateFromSignedFile($file.FullName).Subject
+            $vendor = if ($subject -match '(?:^|,\s*)O=("[^"]*"|[^,]*)') { $Matches[1].Trim('"') } else { '' }
+            if ($vendor -notin $allowedVendors) {
+              Write-Host "::error file=$($file.Name)::Signed by '$subject' -- vendor '$vendor' is not on the allowlist."
+              $failed = $true
+              continue
+            }
+            Write-Host "$($file.Name): embedded signature valid, vendor $vendor"
+          }
+
+          if ($failed) { throw "Third-party signature check failed. Nothing was published." }
+          Write-Host "All $checked third-party PE files carry a valid signature from an allowed vendor."
+
+      # The payload contract: one catalog, signed like our binaries, holding the SHA-256 of every
+      # other file in the ZIP. Installer, OOBE script and self-updater accept a payload only when
+      # this catalog is ours and matches the extracted files exactly. Created AFTER signing and
+      # BEFORE the ZIP -- the same load-bearing position as the signing step above.
+      - name: Create and sign payload catalog
+        shell: pwsh
+        env:
+          DOTNET_ROLL_FORWARD: LatestMajor
+        run: |
+          $ErrorActionPreference = 'Stop'
+          $stagingDir = 'staging/agent'
+          $catalog = Join-Path $env:RUNNER_TEMP $env:PAYLOAD_CATALOG
+          if (Test-Path (Join-Path $stagingDir $env:PAYLOAD_CATALOG)) {
+            throw "$stagingDir already contains $env:PAYLOAD_CATALOG."
+          }
+
+          New-FileCatalog -Path $stagingDir -CatalogFilePath $catalog -CatalogVersion 2.0 | Out-Null
+
+          & azuresigntool.exe sign --verbose `
+            --azure-key-vault-url "$env:CODE_SIGN_KEY_VAULT_URL" `
+            --azure-key-vault-accesstoken "$env:CODE_SIGN_AKV_ACCESS_TOKEN" `
+            --azure-key-vault-certificate "$env:CODE_SIGN_CERTIFICATE_NAME" `
+            --timestamp-rfc3161 "$env:CODE_SIGN_TIMESTAMP_URL" `
+            --file-digest sha256 `
+            --description "Autopilot Monitor Agent" `
+            --description-url "$env:WWW_BASE_URL" `
+            $catalog
+          if ($LASTEXITCODE -ne 0) {
+            throw "azuresigntool failed on the payload catalog with exit code $LASTEXITCODE -- aborting before anything is published."
+          }
+
+          Copy-Item -Path $catalog -Destination $stagingDir
+          $count = (Test-FileCatalog -Path $stagingDir -CatalogFilePath $catalog -FilesToSkip $env:PAYLOAD_CATALOG -Detailed).CatalogItems.Count
+          Write-Host "Payload catalog signed: $count files"
+
+      - name: Create release artifact
+        id: artifact
+        shell: pwsh
+        run: |
+          $version = $env:AGENT_VERSION
+          $stagingDir = "staging/agent"
+          $manifest = Get-Content "$env:AGENT_OUTPUT/$env:MANIFEST_NAME" -Raw | ConvertFrom-Json
+
           $versionedZip = "AutopilotMonitor-Agent-v$version.zip"
           Compress-Archive -Path "$stagingDir/*" -DestinationPath $versionedZip
           Write-Host "Created GitHub-Release artifact: $versionedZip"
@@ -365,17 +613,66 @@ jobs:
           }
           Write-Host "All $($files.Count) binaries signed with certificate $($thumbprints.Keys -join ', ')"
 
+      # The devices' check, run on the FINAL ZIP rather than the staging folder: whatever the
+      # archiver did (Compress-Archive skips hidden files) fails here instead of on a device.
+      # Windows PowerShell 5.1 and the installer's own functions (dot-sourced; its entry guard
+      # keeps that side-effect free) -- the very code the installer and the OOBE go-script run.
+      - name: Verify payload catalog (device check)
+        shell: powershell
+        env:
+          RELEASE_ZIP: ${{ steps.artifact.outputs.versioned_zip_release }}
+        run: |
+          $ErrorActionPreference = 'Stop'
+          Add-Type -AssemblyName System.IO.Compression.FileSystem
+
+          $payload = Join-Path $env:RUNNER_TEMP 'verify-payload'
+          if (Test-Path $payload) { throw "$payload already exists." }
+          [System.IO.Compression.ZipFile]::ExtractToDirectory((Resolve-Path $env:RELEASE_ZIP).Path, $payload)
+
+          . (Resolve-Path $env:BOOTSTRAP_SOURCE).Path
+          if (-not [string]::IsNullOrWhiteSpace($env:CODE_SIGN_EXPECTED_SUBJECT) -and $ExpectedPublisher -ne $env:CODE_SIGN_EXPECTED_SUBJECT) {
+            throw "Installer pin '$ExpectedPublisher' differs from CODE_SIGN_EXPECTED_SUBJECT '$env:CODE_SIGN_EXPECTED_SUBJECT'."
+          }
+          if ($AgentCatalogName -ne $env:PAYLOAD_CATALOG) {
+            throw "Installer catalog name '$AgentCatalogName' differs from PAYLOAD_CATALOG '$env:PAYLOAD_CATALOG'."
+          }
+
+          # Assign first, then flatten: the function returns its list as ONE object (comma
+          # operator), so @(Get-AgentPayloadRejections ...) would nest it inside another array.
+          $problems = Get-AgentPayloadRejections -Path $payload -ExpectedPublisher $ExpectedPublisher -CatalogName $AgentCatalogName
+          $problems = @($problems)
+
+          # Release-only checks a device need not repeat: a timestamp (without it the signature
+          # dies with the certificate) and the same certificate as the binaries.
+          $signature = Get-AuthenticodeSignature (Join-Path $payload $AgentCatalogName)
+          if (-not $signature.TimeStamperCertificate) {
+            $problems += 'catalog signature carries no RFC-3161 timestamp -- it would expire with the certificate'
+          }
+          $exeSigner = (Get-AuthenticodeSignature (Join-Path $payload 'AutopilotMonitor.Agent.exe')).SignerCertificate
+          if ($signature.SignerCertificate -and $exeSigner -and $signature.SignerCertificate.Thumbprint -ne $exeSigner.Thumbprint) {
+            $problems += 'catalog and agent exe carry different signing certificates'
+          }
+
+          if ($problems.Count -gt 0) {
+            $problems | ForEach-Object { Write-Host "::error::$_" }
+            throw "Payload catalog verification failed. Nothing was published."
+          }
+          Write-Host "Payload catalog verified on the final ZIP with the installer's own check, signer $($signature.SignerCertificate.Subject)"
+
       - name: Upload to Azure Blob Storage
         if: inputs.upload_blob
         shell: pwsh
+        env:
+          ZIP_SHA256: ${{ steps.artifact.outputs.sha256 }}
+          EXE_SHA256: ${{ steps.artifact.outputs.exe_sha256 }}
         run: |
-          $containerUrl = "https://autopilotmonitoreu.blob.core.windows.net/agent"
+          $containerUrl = $env:AGENT_CONTAINER_URL
           $authHeaders = & ./.github/scripts/Get-StorageAuthHeaders.ps1
           # Legacy mirror (transition until customers moved to the download.-alias):
           # blobs are ALSO pushed to the old account so the bootstrap URL baked into
           # customers' Intune scripts keeps serving current bits. Fail-soft -- a
           # legacy hiccup must never fail a release. Counter lives ONLY on eu.
-          $legacyContainerUrl = "https://autopilotmonitor.blob.core.windows.net/agent"
+          $legacyContainerUrl = $env:LEGACY_AGENT_CONTAINER_URL
           $legacySasToken = "${{ secrets.AZURE_BLOB_AGENT_LEGACY_SAS_TOKEN }}".TrimStart('?')
 
           function Send-Blob {
@@ -405,6 +702,18 @@ jobs:
             }
           }
 
+          # Dev channel: lab blobs only, under the names the dev renders of the bootstrap scripts
+          # read. Nothing else is touched -- no versioned or stable line, no AdminConfiguration, no
+          # scripts; the channel check in `reserve` already refused publish_as_stable/create_release.
+          if ($env:CHANNEL -eq 'dev') {
+            Send-Blob -Local $env:VERSIONED_ZIP -BlobName $env:DEV_VERSIONED_ZIP -ContentType "application/zip"
+            Send-Blob -Local $env:MANIFEST_NAME -BlobName $env:DEV_MANIFEST -ContentType "application/json"
+            Send-Blob -Local $env:VERSIONED_ZIP -BlobName $env:DEV_STABLE_ZIP -ContentType "application/zip"
+            Send-Blob -Local $env:MANIFEST_NAME -BlobName $env:DEV_STABLE_MANIFEST -ContentType "application/json"
+            Write-Host "DEV CHANNEL: lab blobs written, nothing else touched."
+            exit 0
+          }
+
           # Versioned namespace (always)
           Send-Blob -Local $env:VERSIONED_ZIP -BlobName $env:VERSIONED_ZIP -ContentType "application/zip"
           Send-Blob -Local $env:MANIFEST_NAME -BlobName $env:MANIFEST_NAME -ContentType "application/json"
@@ -426,7 +735,6 @@ jobs:
               -SigningToken $env:CODE_SIGN_AKV_ACCESS_TOKEN `
               -KeyVaultUrl $env:CODE_SIGN_KEY_VAULT_URL `
               -CertificateName $env:CODE_SIGN_CERTIFICATE_NAME `
-              -TimestampUrl $env:CODE_SIGN_TIMESTAMP_URL `
               -ExpectedPublisher $env:CODE_SIGN_EXPECTED_SUBJECT
           }
 
@@ -437,10 +745,10 @@ jobs:
           if ('${{ inputs.publish_as_stable }}' -ne 'true') {
             Write-Host "SKIPPED: AdminConfiguration update (versioned-only run -- stable namespace unchanged)"
           } else {
-            $sha256 = "${{ steps.artifact.outputs.sha256 }}"
-            $exeSha256 = "${{ steps.artifact.outputs.exe_sha256 }}"
-            $agentVersion = "${{ steps.version.outputs.version }}"
-            $tableUrl = "https://autopilotmonitoreu.table.core.windows.net/AdminConfiguration(PartitionKey='GlobalConfig',RowKey='config')"
+            $sha256 = $env:ZIP_SHA256
+            $exeSha256 = $env:EXE_SHA256
+            $agentVersion = $env:AGENT_VERSION
+            $tableUrl = $env:ADMIN_CONFIG_URL
             $tableBody = @{
               LatestAgentV2Version    = $agentVersion
               LatestAgentV2Sha256     = $sha256
@@ -464,7 +772,7 @@ jobs:
       # Skipped on runs that publish nothing -- no log entry for bytes nobody receives.
       - name: Attest build provenance
         if: inputs.upload_blob || inputs.create_release
-        uses: actions/attest-build-provenance@4d101475d8b20a2381f78447822ac1eab6504dd8 # v4
+        uses: actions/attest-build-provenance@4d101475d8b20a2381f78447822ac1eab6504dd8 # v4.2.2
         with:
           subject-path: |
             ${{ steps.artifact.outputs.versioned_zip_release }}
@@ -472,13 +780,13 @@ jobs:
             staging/agent/AutopilotMonitor.SummaryDialog.exe
 
       - name: Upload build artifact
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
         with:
-          name: AutopilotMonitor-Agent-v${{ steps.version.outputs.version }}
+          name: AutopilotMonitor-Agent-v${{ needs.reserve.outputs.version }}
           path: ${{ steps.artifact.outputs.versioned_zip_release }}
 
       - name: Upload version metadata
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
         with:
           name: version-metadata
           path: ${{ env.MANIFEST_NAME }}
@@ -497,30 +805,30 @@ jobs:
         shell: pwsh
         env:
           GH_TOKEN: ${{ github.token }}
-          RELEASE_TAG: ${{ steps.version.outputs.tag }}
-          RELEASE_NAME: Autopilot Monitor Agent ${{ steps.version.outputs.version }}
+          RELEASE_TAG: ${{ needs.reserve.outputs.tag }}
+          RELEASE_NAME: Autopilot Monitor Agent ${{ needs.reserve.outputs.version }}
           RELEASE_PRERELEASE: ${{ inputs.prerelease }}
           RELEASE_ZIP: ${{ steps.artifact.outputs.versioned_zip_release }}
           RELEASE_NOTES: ${{ inputs.release_notes }}
           RELEASE_BODY_TAIL: |
             ### Version Info
-            - **Version:** ${{ steps.version.outputs.version }}
-            - **Build Number:** ${{ steps.version.outputs.build_number }}
+            - **Version:** ${{ needs.reserve.outputs.version }}
+            - **Build Number:** ${{ needs.reserve.outputs.build_number }}
             - **SHA-256:** `${{ steps.artifact.outputs.sha256 }}`
             - **Built from:** ${{ github.sha }}
             - **Published as stable:** ${{ inputs.publish_as_stable }}
 
             ### Provenance
-            The executables are Authenticode-signed and timestamped. The ZIP and both executables carry a [GitHub Artifact Attestation](https://github.com/okieselbach/AutopilotMonitor/attestations) (Sigstore keyless). Verify:
+            The executables are Authenticode-signed and timestamped. Every file in the ZIP is listed in `AutopilotMonitor.Agent.cat`, a catalog signed with the same certificate, and third-party libraries carry their vendor's signature. The ZIP and both executables carry a [GitHub Artifact Attestation](https://github.com/okieselbach/AutopilotMonitor/attestations) (Sigstore keyless). Verify:
             ```
-            gh attestation verify AutopilotMonitor-Agent-v${{ steps.version.outputs.version }}.zip --repo okieselbach/AutopilotMonitor
+            gh attestation verify AutopilotMonitor-Agent-v${{ needs.reserve.outputs.version }}.zip --repo okieselbach/AutopilotMonitor
             ```
         run: |
           $ErrorActionPreference = 'Stop'
 
           $notes = "$env:RELEASE_NOTES".Trim()
           if ($notes -eq '') {
-            $notes = "Minor enhancements and bug fixes. See the [Agent Changelog](https://docs.autopilotmonitor.com/changelog/agent-changelog) for details.`n<!-- release-notes: none -->"
+            $notes = "Minor enhancements and bug fixes. See the [Agent Changelog]($env:DOCS_BASE_URL/changelog/agent-changelog) for details.`n<!-- release-notes: none -->"
           }
 
           # Previous release for the compare link. Fail-soft: the first release has none,

.github/workflows/build-msi.yml

@@ -11,39 +11,54 @@ on:
   workflow_dispatch:
     inputs:
       upload_blob:
-        description: 'Upload to blob (customer-facing via download.autopilotmonitor.com/agent/)'
+        description: 'Upload to blob (customer-facing through the download alias)'
         required: true
         type: boolean
         default: true
 
+# Every job states its own permissions. This default is what the build job runs with: wix is
+# third-party code and must never hold more than a read-only token.
 permissions:
   contents: read
-  id-token: write      # keyless Sigstore signing (artifact attestation)
-  attestations: write  # store the provenance attestation
 
 env:
   MSI_PROJECT_DIR: src/Agent/AutopilotMonitor.BootstrapMsi
   RUNNER_SCRIPT: scripts/Bootstrap/Start-AutopilotMonitor.ps1
   STABLE_MSI: AutopilotMonitor-Bootstrap.msi
-  # Authenticode via Azure Key Vault; identity and vault come from repository variables.
+  # Authenticode via Azure Key Vault; identity and vault come from repository variables, the
+  # timestamp server (CODE_SIGN_TIMESTAMP_URL) from the endpoint registry like every host here.
   # CODE_SIGN_EXPECTED_SUBJECT is optional and pins the publisher in the verify steps.
   CODE_SIGN_CLIENT_ID: ${{ vars.CODE_SIGN_CLIENT_ID }}
   CODE_SIGN_CERTIFICATE_NAME: ${{ vars.CODE_SIGN_CERTIFICATE_NAME }}
   CODE_SIGN_KEY_VAULT_URL: ${{ vars.CODE_SIGN_KEY_VAULT_URL }}
   CODE_SIGN_EXPECTED_SUBJECT: ${{ vars.CODE_SIGN_EXPECTED_SUBJECT }}
-  CODE_SIGN_TIMESTAMP_URL: http://timestamp.acs.microsoft.com
 
 jobs:
-  build:
+  # Signs the loader BEFORE the build: wix embeds the file as it finds it, so a loader signed
+  # afterwards would leave the copy inside the MSI unsigned. Devices under an AllSigned policy
+  # need the embedded loader to carry a signature of its own.
+  #
+  # Trust zone: only this repository's scripts, pinned actions and the pinned signing tool run
+  # here. No `environment:` yet: the code-signing credential is still bound to the ref, and a
+  # job carries exactly one OIDC subject.
+  sign-loader:
+    name: Sign bootstrap loader
     runs-on: windows-latest
-    name: Build Bootstrap MSI
+    permissions:
+      contents: read
+      id-token: write   # code-signing login
 
     steps:
       - name: Checkout repository
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
+      # Every host in this workflow comes from the one endpoint registry.
+      - name: Load endpoints
+        shell: pwsh
+        run: ./.github/scripts/Get-Endpoints.ps1 -ExportToGitHubEnv
+
       # The loader executes as SYSTEM on every enrolling device — a parse error or
       # non-ASCII mojibake (PS 5.1 codepage roulette) must fail the build, not the fleet.
       # bootstrap-script-gates.yml runs the same checks plus Pester on pushes; this job
@@ -60,6 +75,151 @@ jobs:
           }
           Write-Host "Runner script parse + ASCII check passed"
 
+      # Every published MSI is signed; a missing variable stops the run before the build.
+      - name: Check code-signing configuration
+        shell: pwsh
+        run: |
+          foreach ($name in 'CODE_SIGN_CLIENT_ID', 'CODE_SIGN_CERTIFICATE_NAME', 'CODE_SIGN_KEY_VAULT_URL') {
+            if ([string]::IsNullOrWhiteSpace([Environment]::GetEnvironmentVariable($name))) {
+              throw "Repository variable $name is not set. Every published MSI must be Authenticode-signed."
+            }
+          }
+
+      - name: Login to Azure (code-signing identity)
+        uses: azure/login@a641126d1b8aa4d1fa005f4f92df94a3a4c4c906 # v3.1.0
+        with:
+          client-id: ${{ env.CODE_SIGN_CLIENT_ID }}
+          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
+          allow-no-subscriptions: true
+
+      - name: Acquire Key Vault token for signing
+        shell: pwsh
+        run: |
+          $ErrorActionPreference = 'Stop'
+          $token = az account get-access-token --scope https://vault.azure.net/.default --query accessToken --output tsv
+          if ($LASTEXITCODE -ne 0 -or [string]::IsNullOrWhiteSpace($token)) {
+            throw "Could not acquire a Key Vault token for the code-signing identity."
+          }
+          Write-Host "::add-mask::$token"
+          "CODE_SIGN_AKV_ACCESS_TOKEN=$token" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+
+      # Version pinned; roll-forward because the tool targets an older runtime than the SDK.
+      - name: Install AzureSignTool
+        env:
+          DOTNET_ROLL_FORWARD: LatestMajor
+        run: dotnet tool install --no-cache --global AzureSignTool --version 7.0.1
+
+      # A missing timestamp fails the run: without it the signature dies with the certificate.
+      - name: Sign bootstrap loader (Authenticode)
+        shell: pwsh
+        env:
+          DOTNET_ROLL_FORWARD: LatestMajor
+        run: |
+          $ErrorActionPreference = 'Stop'
+          & azuresigntool.exe sign --verbose `
+            --azure-key-vault-url "$env:CODE_SIGN_KEY_VAULT_URL" `
+            --azure-key-vault-accesstoken "$env:CODE_SIGN_AKV_ACCESS_TOKEN" `
+            --azure-key-vault-certificate "$env:CODE_SIGN_CERTIFICATE_NAME" `
+            --timestamp-rfc3161 "$env:CODE_SIGN_TIMESTAMP_URL" `
+            --file-digest sha256 `
+            --description 'Autopilot Monitor bootstrap runner' `
+            --description-url $env:WWW_BASE_URL `
+            $env:RUNNER_SCRIPT
+          if ($LASTEXITCODE -ne 0) { throw "azuresigntool failed with exit code $LASTEXITCODE" }
+
+          $sig = Get-AuthenticodeSignature -FilePath $env:RUNNER_SCRIPT
+          if ($sig.Status -ne 'Valid') { throw "Signature on $env:RUNNER_SCRIPT is '$($sig.Status)', expected 'Valid'." }
+          if (-not $sig.TimeStamperCertificate) { throw "Signature on $env:RUNNER_SCRIPT carries no timestamp." }
+          if ($env:CODE_SIGN_EXPECTED_SUBJECT -and $sig.SignerCertificate.Subject -notlike $env:CODE_SIGN_EXPECTED_SUBJECT) {
+            throw "$env:RUNNER_SCRIPT was signed by '$($sig.SignerCertificate.Subject)', expected '$env:CODE_SIGN_EXPECTED_SUBJECT'."
+          }
+          Write-Host "$env:RUNNER_SCRIPT : Valid, timestamped, signer $($sig.SignerCertificate.Subject.Split(',')[0])"
+
+      - name: Upload signed loader
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
+        with:
+          name: bootstrap-loader-signed
+          path: ${{ env.RUNNER_SCRIPT }}
+          if-no-files-found: error
+          retention-days: 1
+
+  # Build zone: no login, no secret, no id-token. Its only output is the artifact; the
+  # release job treats that as data and never builds.
+  build:
+    name: Build MSI
+    needs: sign-loader
+    runs-on: windows-latest
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
+
+      - name: Download signed loader
+        uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
+        with:
+          name: bootstrap-loader-signed
+          path: ${{ runner.temp }}/loader
+
+      # The signed copy replaces the checkout's unsigned one before wix embeds it.
+      - name: Put the signed loader in place
+        shell: pwsh
+        run: |
+          $ErrorActionPreference = 'Stop'
+          Copy-Item -Path (Join-Path $env:RUNNER_TEMP 'loader/Start-AutopilotMonitor.ps1') -Destination $env:RUNNER_SCRIPT -Force
+          $sig = Get-AuthenticodeSignature -FilePath $env:RUNNER_SCRIPT
+          if ($sig.Status -ne 'Valid') { throw "Loader to embed is '$($sig.Status)', expected 'Valid'." }
+          Write-Host "Embedding the signed loader ($((Get-Item $env:RUNNER_SCRIPT).Length) bytes)"
+
+      # Exact version, not a "5.*" range: wix writes the bytes of the MSI that customers
+      # install, so a compromised or simply broken point release must not flow into a build
+      # unnoticed. Bumping is a reviewed commit.
+      - name: Install WiX toolset
+        run: dotnet tool install --global wix --version 5.0.2
+
+      - name: Build MSI
+        shell: pwsh
+        run: |
+          # wix resolves File/@Source relative to the working directory, not the wxs —
+          # build from inside the project dir (same as the local build documented in the wxs).
+          $out = Join-Path (Get-Location) $env:STABLE_MSI
+          Push-Location $env:MSI_PROJECT_DIR
+          wix build Package.wxs -arch x64 -o $out
+          Pop-Location
+          if (-not (Test-Path $env:STABLE_MSI)) { Write-Error "wix build produced no MSI"; exit 1 }
+          Write-Host "Built $env:STABLE_MSI ($([math]::Round((Get-Item $env:STABLE_MSI).Length / 1KB)) KB)"
+
+      - name: Upload unsigned MSI
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
+        with:
+          name: bootstrap-msi-unsigned
+          path: ${{ env.STABLE_MSI }}
+          if-no-files-found: error
+          retention-days: 1
+
+  # Trust zone: signs, verifies and publishes the MSI as data; never builds. No
+  # `environment:` yet, for the same reason as sign-loader.
+  release:
+    name: Sign and publish MSI
+    needs: build
+    runs-on: windows-latest
+    permissions:
+      contents: read
+      id-token: write      # code-signing and storage logins, keyless Sigstore signing
+      attestations: write  # store the provenance attestation
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
+
+      - name: Load endpoints
+        shell: pwsh
+        run: ./.github/scripts/Get-Endpoints.ps1 -ExportToGitHubEnv
+
+      # Read from this job's own checkout, never handed over by the build job.
       - name: Read MSI version from Package.wxs
         id: version
         shell: pwsh
@@ -73,20 +233,19 @@ jobs:
           Write-Host "MSI version (from Package.wxs): $version"
           "version=$version" >> $env:GITHUB_OUTPUT
 
-      # Every published MSI is signed; a missing variable stops the run before the build.
-      - name: Check code-signing configuration
+      - name: Download unsigned MSI
+        uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
+        with:
+          name: bootstrap-msi-unsigned
+          path: ${{ runner.temp }}/msi
+
+      # Only the expected file leaves the download folder.
+      - name: Take the MSI from the build
         shell: pwsh
-        run: |
-          foreach ($name in 'CODE_SIGN_CLIENT_ID', 'CODE_SIGN_CERTIFICATE_NAME', 'CODE_SIGN_KEY_VAULT_URL') {
-            if ([string]::IsNullOrWhiteSpace([Environment]::GetEnvironmentVariable($name))) {
-              throw "Repository variable $name is not set. Every published MSI must be Authenticode-signed."
-            }
-          }
+        run: Copy-Item -Path (Join-Path $env:RUNNER_TEMP "msi/$env:STABLE_MSI") -Destination $env:STABLE_MSI -ErrorAction Stop
 
       # Logs in FIRST and only to fetch the Key Vault token: azure/login sets the default az
-      # account, so the storage login below has to be the last one. Both federated credentials
-      # are bound to the ref, so this job must never be given an `environment:` -- that would
-      # change the OIDC subject claim and break both logins.
+      # account, so the storage login below has to be the last one.
       - name: Login to Azure (code-signing identity)
         uses: azure/login@a641126d1b8aa4d1fa005f4f92df94a3a4c4c906 # v3.1.0
         with:
@@ -113,50 +272,12 @@ jobs:
           tenant-id: ${{ secrets.AZURE_TENANT_ID }}
           subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}
 
-      # Exact version, not a "5.*" range: wix writes the bytes of the MSI that customers
-      # install, so a compromised or simply broken point release must not flow into a build
-      # unnoticed. Bumping is a reviewed commit.
-      - name: Install WiX toolset
-        run: dotnet tool install --global wix --version 5.0.2
-
       # Version pinned; roll-forward because the tool targets an older runtime than the SDK.
       - name: Install AzureSignTool
         env:
           DOTNET_ROLL_FORWARD: LatestMajor
         run: dotnet tool install --no-cache --global AzureSignTool --version 7.0.1
 
-      # BEFORE the build: wix embeds the file as it finds it, so a loader signed afterwards
-      # would leave the copy inside the MSI unsigned. Devices under an AllSigned policy need
-      # the embedded loader to carry a signature of its own.
-      - name: Sign bootstrap loader (Authenticode)
-        shell: pwsh
-        env:
-          DOTNET_ROLL_FORWARD: LatestMajor
-        run: |
-          $ErrorActionPreference = 'Stop'
-          & azuresigntool.exe sign --verbose `
-            --azure-key-vault-url "$env:CODE_SIGN_KEY_VAULT_URL" `
-            --azure-key-vault-accesstoken "$env:CODE_SIGN_AKV_ACCESS_TOKEN" `
-            --azure-key-vault-certificate "$env:CODE_SIGN_CERTIFICATE_NAME" `
-            --timestamp-rfc3161 "$env:CODE_SIGN_TIMESTAMP_URL" `
-            --file-digest sha256 `
-            --description 'Autopilot Monitor bootstrap runner' `
-            --description-url 'https://www.autopilotmonitor.com' `
-            $env:RUNNER_SCRIPT
-          if ($LASTEXITCODE -ne 0) { throw "azuresigntool failed with exit code $LASTEXITCODE" }
-
-      - name: Build MSI
-        shell: pwsh
-        run: |
-          # wix resolves File/@Source relative to the working directory, not the wxs —
-          # build from inside the project dir (same as the local build documented in the wxs).
-          $out = Join-Path (Get-Location) $env:STABLE_MSI
-          Push-Location $env:MSI_PROJECT_DIR
-          wix build Package.wxs -arch x64 -o $out
-          Pop-Location
-          if (-not (Test-Path $env:STABLE_MSI)) { Write-Error "wix build produced no MSI"; exit 1 }
-          Write-Host "Built $env:STABLE_MSI ($([math]::Round((Get-Item $env:STABLE_MSI).Length / 1KB)) KB)"
-
       # Signing rewrites the MSI, so it runs before the hash and the upload. A missing
       # timestamp fails the run: without it the signature dies with the certificate.
       - name: Sign MSI (Authenticode)
@@ -172,30 +293,30 @@ jobs:
             --timestamp-rfc3161 "$env:CODE_SIGN_TIMESTAMP_URL" `
             --file-digest sha256 `
             --description 'Autopilot Monitor Bootstrap' `
-            --description-url 'https://www.autopilotmonitor.com' `
+            --description-url $env:WWW_BASE_URL `
             $env:STABLE_MSI
           if ($LASTEXITCODE -ne 0) { throw "azuresigntool failed with exit code $LASTEXITCODE" }
 
-          # The MSI and the loader it carries -- verified here, before anything is published.
-          foreach ($file in @($env:STABLE_MSI, $env:RUNNER_SCRIPT)) {
-            $sig = Get-AuthenticodeSignature -FilePath $file
-            if ($sig.Status -ne 'Valid') { throw "Signature on $file is '$($sig.Status)', expected 'Valid'." }
-            if (-not $sig.TimeStamperCertificate) { throw "Signature on $file carries no timestamp." }
-            if ($env:CODE_SIGN_EXPECTED_SUBJECT -and $sig.SignerCertificate.Subject -notlike $env:CODE_SIGN_EXPECTED_SUBJECT) {
-              throw "$file was signed by '$($sig.SignerCertificate.Subject)', expected '$env:CODE_SIGN_EXPECTED_SUBJECT'."
-            }
-            Write-Host "$file : Valid, timestamped, signer $($sig.SignerCertificate.Subject.Split(',')[0])"
+          # The loader was verified in sign-loader; here the MSI, before anything is published.
+          $sig = Get-AuthenticodeSignature -FilePath $env:STABLE_MSI
+          if ($sig.Status -ne 'Valid') { throw "Signature on $env:STABLE_MSI is '$($sig.Status)', expected 'Valid'." }
+          if (-not $sig.TimeStamperCertificate) { throw "Signature on $env:STABLE_MSI carries no timestamp." }
+          if ($env:CODE_SIGN_EXPECTED_SUBJECT -and $sig.SignerCertificate.Subject -notlike $env:CODE_SIGN_EXPECTED_SUBJECT) {
+            throw "$env:STABLE_MSI was signed by '$($sig.SignerCertificate.Subject)', expected '$env:CODE_SIGN_EXPECTED_SUBJECT'."
           }
+          Write-Host "$env:STABLE_MSI : Valid, timestamped, signer $($sig.SignerCertificate.Subject.Split(',')[0])"
 
           Write-Host "SHA-256: $((Get-FileHash $env:STABLE_MSI -Algorithm SHA256).Hash.ToLowerInvariant())"
 
       - name: Upload to Azure Blob Storage
         if: inputs.upload_blob
         shell: pwsh
+        env:
+          MSI_VERSION: ${{ steps.version.outputs.version }}
         run: |
-          $containerUrl = "https://autopilotmonitoreu.blob.core.windows.net/agent"
+          $containerUrl = $env:AGENT_CONTAINER_URL
           $authHeaders = & ./.github/scripts/Get-StorageAuthHeaders.ps1
-          $version = "${{ steps.version.outputs.version }}"
+          $version = $env:MSI_VERSION
 
           # Same Send-Blob contract as build-agent.yml: no-cache on every blob because the
           # stable name rotates in place behind the download alias. No legacy mirror — the
@@ -226,26 +347,26 @@ jobs:
         shell: pwsh
         run: |
           $localSize = (Get-Item $env:STABLE_MSI).Length
-          $resp = Invoke-WebRequest -Uri "https://download.autopilotmonitor.com/agent/$env:STABLE_MSI" -Method Head -UseBasicParsing
+          $resp = Invoke-WebRequest -Uri "$env:DOWNLOAD_ALIAS_URL/$env:STABLE_MSI" -Method Head -UseBasicParsing
           if ($resp.StatusCode -ne 200) { Write-Error "Alias HEAD returned $($resp.StatusCode)"; exit 1 }
           $remoteSize = [long]$resp.Headers['Content-Length'][0]
           if ($remoteSize -ne $localSize) {
             Write-Error "Size mismatch: alias serves $remoteSize bytes, local MSI is $localSize bytes"
             exit 1
           }
-          Write-Host "Alias verified: https://download.autopilotmonitor.com/agent/$env:STABLE_MSI ($remoteSize bytes)"
+          Write-Host "Alias verified: $env:DOWNLOAD_ALIAS_URL/$env:STABLE_MSI ($remoteSize bytes)"
 
       # Sigstore keyless provenance, same as the agent ZIP; covers the signed bytes. Skipped
       # on runs that publish nothing -- no log entry for bytes nobody receives. Verify:
       # gh attestation verify AutopilotMonitor-Bootstrap.msi --repo okieselbach/AutopilotMonitor
       - name: Attest build provenance (MSI)
         if: inputs.upload_blob
-        uses: actions/attest-build-provenance@4d101475d8b20a2381f78447822ac1eab6504dd8 # v4
+        uses: actions/attest-build-provenance@4d101475d8b20a2381f78447822ac1eab6504dd8 # v4.2.2
         with:
           subject-path: ${{ env.STABLE_MSI }}
 
       - name: Upload build artifact
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
         with:
           name: AutopilotMonitor-Bootstrap-v${{ steps.version.outputs.version }}
           path: ${{ env.STABLE_MSI }}

.github/workflows/ci.yml

@@ -1,8 +1,8 @@
 name: CI
 
 # Tests, typechecks and drift-checks every push to main and every pull request:
-# the .NET suites, the web and MCP suites, and a guard that the generated rule
-# files still match rules/guardrails.json.
+# the .NET suites, the web and MCP suites, a guard that the generated rule
+# files still match rules/guardrails.json, and the duplicate-code check.
 #
 # The deploy workflows build but never test, so this is the only place the suites
 # actually run. It does not gate anything — main is deliberately unprotected, so a
@@ -46,7 +46,6 @@ on:
       - '.github/workflows/deploy-mcp.yml'
       - '.github/workflows/deploy-web.yml'
       - '.github/workflows/indexnow-notify.yml'
-      - '.github/workflows/oidc-login-test.yml'
       - '.github/workflows/pester.yml'
       - '.github/workflows/publish-scripts.yml'
   pull_request:
@@ -80,7 +79,6 @@ on:
       - '.github/workflows/deploy-mcp.yml'
       - '.github/workflows/deploy-web.yml'
       - '.github/workflows/indexnow-notify.yml'
-      - '.github/workflows/oidc-login-test.yml'
       - '.github/workflows/pester.yml'
       - '.github/workflows/publish-scripts.yml'
   workflow_dispatch:
@@ -98,11 +96,11 @@ jobs:
     # windows-latest is required: the agent test projects target net48.
     runs-on: windows-latest
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
-      - uses: actions/setup-dotnet@a98b56852c35b8e3190ac28c8c2271da59106c68 # v6
+      - uses: actions/setup-dotnet@a98b56852c35b8e3190ac28c8c2271da59106c68 # v6.0.0
         with:
           dotnet-version: '10.0.x'
 
@@ -124,6 +122,9 @@ jobs:
       - name: Summary Dialog tests
         run: dotnet test src/Agent/AutopilotMonitor.SummaryDialog.Tests --configuration Release --no-build --verbosity quiet --logger "console;verbosity=minimal"
 
+      - name: Push tests
+        run: dotnet test src/Push/AutopilotMonitor.Push.Tests --configuration Release --no-build --verbosity quiet --logger "console;verbosity=minimal"
+
   web:
     name: Web (typecheck + tests)
     runs-on: ubuntu-latest
@@ -131,11 +132,11 @@ jobs:
       run:
         working-directory: src/Web/autopilot-monitor-web
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
           node-version: '22'
           cache: npm
@@ -157,6 +158,12 @@ jobs:
       - name: Lint
         run: npm run lint -- --max-warnings 0
 
+      # Unused files, exports and dependencies (knip.json). Its own script on
+      # purpose: chained into `lint`, the `-- --max-warnings 0` above would be
+      # handed to knip and ESLint would lose its cap.
+      - name: Unused code
+        run: npm run knip
+
       - name: Tests
         run: npm test
 
@@ -167,11 +174,11 @@ jobs:
       run:
         working-directory: src/McpServer/autopilot-monitor-mcp
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
           node-version: '22'
           cache: npm
@@ -179,9 +186,19 @@ jobs:
 
       - run: npm ci
 
+      # tsconfig.json covers sources, tests and scripts; the build emits via
+      # tsconfig.build.json.
       - name: Typecheck
         run: npx tsc --noEmit
 
+      # Same rule as the web job: any warning fails — fix new findings, never
+      # raise the cap.
+      - name: Lint
+        run: npm run lint -- --max-warnings 0
+
+      - name: Unused code
+        run: npm run knip
+
       # The vector-search tests tokenize with the real all-MiniLM-L6-v2 model
       # (~23 MB tokenizer + ONNX weights). transformers.js caches under
       # node_modules by default, which npm ci recreates — so every CI run
@@ -189,7 +206,7 @@ jobs:
       # (run 31252869254). Persist the model across runs instead; HF_CACHE_DIR
       # is honored by vector-search-provider.ts.
       - name: Cache embedding model
-        uses: actions/cache@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6
+        uses: actions/cache@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
         with:
           path: ${{ github.workspace }}/.hf-cache
           key: hf-model-Xenova-all-MiniLM-L6-v2
@@ -203,11 +220,11 @@ jobs:
     name: Generated guardrails match rules/guardrails.json
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
           node-version: '22'
 
@@ -226,11 +243,11 @@ jobs:
     name: Generated wire types match shared-manifests.json
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
           node-version: '22'
 
@@ -244,8 +261,33 @@ jobs:
 
       - name: Fail if the generated files drifted
         run: |
-          if ! git diff --quiet -- src/Web/autopilot-monitor-web/utils/shared-manifests.generated.ts src/Web/autopilot-monitor-web/utils/wire-types.generated.ts src/McpServer/autopilot-monitor-mcp/src/generated/wire-types.generated.ts src/McpServer/autopilot-monitor-mcp/src/generated/wire-vocabularies.generated.ts; then
+          if ! git diff --quiet -- src/Web/autopilot-monitor-web/lib/generated/shared-manifests.generated.ts src/Web/autopilot-monitor-web/lib/generated/wire-types.generated.ts src/McpServer/autopilot-monitor-mcp/src/generated/wire-types.generated.ts src/McpServer/autopilot-monitor-mcp/src/generated/wire-vocabularies.generated.ts; then
             echo "::error::Generated wire types are out of date. Run 'npm run generate:manifests' in src/Web/autopilot-monitor-web and commit the result."
             git diff --stat
             exit 1
           fi
+
+  duplicates:
+    name: Duplicate code
+    runs-on: ubuntu-latest
+    defaults:
+      run:
+        working-directory: tools/quality
+    steps:
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
+
+      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
+        with:
+          node-version: '22'
+          cache: npm
+          cache-dependency-path: tools/quality/package-lock.json
+
+      - run: npm ci --ignore-scripts
+
+      # jscpd over the C# and TypeScript sources (.jscpd.json). A clone missing from
+      # jscpd-baseline.json fails, and so does a baseline entry whose clone is gone,
+      # so the baseline only shrinks. Locally: npm run dup in tools/quality.
+      - name: Duplicate check
+        run: npm run dup

.github/workflows/codeql.yml

@@ -28,6 +28,7 @@ on:
       # Every other workflow: none of them changes the C# or TypeScript CodeQL analyzes.
       # paths-ignore cannot negate, so they are listed one by one; codeql.yml itself and
       # .github/codeql/ stay out of the list. A new workflow file is opt-in — add it here too.
+      - '.github/workflows/action-pins.yml'
       - '.github/workflows/bootstrap-script-gates.yml'
       - '.github/workflows/build-agent.yml'
       - '.github/workflows/build-msi.yml'
@@ -37,7 +38,6 @@ on:
       - '.github/workflows/deploy-mcp.yml'
       - '.github/workflows/deploy-web.yml'
       - '.github/workflows/indexnow-notify.yml'
-      - '.github/workflows/oidc-login-test.yml'
       - '.github/workflows/pester.yml'
       - '.github/workflows/publish-scripts.yml'
   pull_request:
@@ -60,6 +60,7 @@ on:
       # Every other workflow: none of them changes the C# or TypeScript CodeQL analyzes.
       # paths-ignore cannot negate, so they are listed one by one; codeql.yml itself and
       # .github/codeql/ stay out of the list. A new workflow file is opt-in — add it here too.
+      - '.github/workflows/action-pins.yml'
       - '.github/workflows/bootstrap-script-gates.yml'
       - '.github/workflows/build-agent.yml'
       - '.github/workflows/build-msi.yml'
@@ -69,7 +70,6 @@ on:
       - '.github/workflows/deploy-mcp.yml'
       - '.github/workflows/deploy-web.yml'
       - '.github/workflows/indexnow-notify.yml'
-      - '.github/workflows/oidc-login-test.yml'
       - '.github/workflows/pester.yml'
       - '.github/workflows/publish-scripts.yml'
   schedule:
@@ -104,11 +104,11 @@ jobs:
             build-mode: none
 
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
-      - uses: actions/setup-dotnet@a98b56852c35b8e3190ac28c8c2271da59106c68 # v6
+      - uses: actions/setup-dotnet@a98b56852c35b8e3190ac28c8c2271da59106c68 # v6.0.0
         if: matrix.language == 'csharp'
         with:
           dotnet-version: '10.0.x'

.github/workflows/combine-rules.yml

@@ -1,7 +1,7 @@
 name: Validate & Combine Rules
 on:
   push:
-    branches: [main, dev]
+    branches: [main]
     paths:
       - 'rules/gather/**'
       - 'rules/analyze/**'
@@ -9,6 +9,7 @@ on:
       - 'rules/schema/**'
       - 'rules/scripts/**'
       - 'rules/guardrails.json'
+      - '.github/workflows/combine-rules.yml'
   pull_request:
     paths:
       - 'rules/gather/**'
@@ -17,73 +18,71 @@ on:
       - 'rules/schema/**'
       - 'rules/scripts/**'
       - 'rules/guardrails.json'
+      - '.github/workflows/combine-rules.yml'
+
+permissions:
+  contents: read
 
 jobs:
-  validate-and-combine:
+  # Pull requests and pushes alike; read-only, because the validator is third-party code.
+  validate:
     runs-on: ubuntu-latest
-    permissions:
-      contents: write
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
 
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
           node-version: '22'
 
-      # Lockfile and no lifecycle scripts: this job holds contents:write and pushes to
-      # main, so an install-time script from a hijacked package would run with a token that
-      # can rewrite the repository. rules/scripts/package-lock.json pins ajv-cli, ajv-formats
-      # and every transitive package (Dependabot keeps it current); ajv is pure JS and
-      # needs no install hooks.
-      - name: Install ajv-cli from the lockfile
+      # Lockfile and no lifecycle scripts: rules/scripts/package-lock.json pins ajv,
+      # ajv-formats and every transitive package (Dependabot keeps it current); ajv is pure
+      # JS and needs no install hooks.
+      - name: Install the rules validator from the lockfile
         working-directory: rules/scripts
-        run: |
-          npm ci --ignore-scripts
-          echo "$PWD/node_modules/.bin" >> "$GITHUB_PATH"
+        run: npm ci --ignore-scripts
 
-      - name: Validate Gather Rules
-        run: |
-          for f in rules/gather/*.json; do
-            echo "Validating $f..."
-            ajv validate -s rules/schema/gather-rule.schema.json -d "$f" --spec=draft2020 -c ajv-formats
-          done
-
-      - name: Validate Analyze Rules
-        run: |
-          for f in rules/analyze/*.json; do
-            echo "Validating $f..."
-            ajv validate -s rules/schema/analyze-rule.schema.json -d "$f" --spec=draft2020 -c ajv-formats
-          done
-
-      - name: Validate IME Log Patterns
-        run: |
-          for f in rules/ime-log-patterns/*.json; do
-            echo "Validating $f..."
-            ajv validate -s rules/schema/ime-log-pattern.schema.json -d "$f" --spec=draft2020 -c ajv-formats
-          done
-
-      # guardrails.json values are baked verbatim into generated TypeScript that is
-      # auto-committed to main and shipped to the portal - the schema forbids control
-      # and line-terminator characters so a "data" PR can never become code.
-      - name: Validate Guardrails
-        run: ajv validate -s rules/schema/guardrails.schema.json -d rules/guardrails.json --spec=draft2020 -c ajv-formats
+      # Gather, analyze and IME log pattern files plus guardrails.json, each against its
+      # schema; the guardrails schema is what keeps a "data" PR from becoming code in the
+      # generated TypeScript (see validate.js).
+      - name: Validate rules and guardrails against their schemas
+        run: node rules/scripts/validate.js
 
       - name: Combine rules into dist/
         run: node rules/scripts/combine.js
 
-      # Plain git instead of a third-party action: this step runs with contents:write on
-      # main, so no code we do not control gets to push. Checkout leaves the job's
-      # GITHUB_TOKEN in the remote credential helper, which is all `git push` needs.
+  # The one job that can push to main, and it installs nothing: combine.js needs only node's
+  # built-ins, so no package code ever runs next to the write token. Keep it that way -- a
+  # dependency in combine.js would need its own job.
+  publish:
+    needs: validate
+    if: github.ref == 'refs/heads/main' && github.event_name == 'push'
+    runs-on: ubuntu-latest
+    permissions:
+      contents: write
+    steps:
+      # Checkout leaves the job's GITHUB_TOKEN in the remote credential helper, which is all
+      # `git push` needs.
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+
+      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
+        with:
+          node-version: '22'
+
+      - name: Combine rules into dist/
+        run: node rules/scripts/combine.js
+
+      # Plain git instead of a third-party action, so no code we do not control gets to push.
       # NOTE: combine.js writes THREE outputs; committing only dist/ would silently discard
       # the two generated TS modules and leave the guardrails-in-sync CI check red.
       - name: Commit generated outputs
-        if: github.ref == 'refs/heads/main' && github.event_name == 'push'
         run: |
           set -euo pipefail
           git config user.name "github-actions[bot]"
           git config user.email "41898282+github-actions[bot]@users.noreply.github.com"
           git add -- rules/dist/ \
-            src/Web/autopilot-monitor-web/utils/guardrails.generated.ts \
+            src/Web/autopilot-monitor-web/lib/generated/guardrails.generated.ts \
             src/McpServer/autopilot-monitor-mcp/src/rule-authoring.generated.ts
           if git diff --cached --quiet; then
             echo "Generated outputs unchanged - nothing to commit"

.github/workflows/deploy-backend.yml

@@ -3,26 +3,41 @@ name: Deploy Backend (Azure Functions)
 on:
   workflow_dispatch:
 
+# Every job states its own permissions. This default is what the build job runs with: restore
+# and build execute third-party code (NuGet MSBuild targets, the Functions worker SDK) and must
+# never hold more than a read-only token.
 permissions:
-  id-token: write   # OIDC federated login (secret-less)
   contents: read
 
+# Hosts (API, versions container) come from .github/scripts/Get-Endpoints.ps1.
 env:
   FUNCTION_APP: autopilotmonitor-api-eu
-  HEALTH_URL: https://autopilotmonitor-api-eu.azurewebsites.net/api/health
-  VERSIONS_CONTAINER: https://autopilotmonitoreu.blob.core.windows.net/versions
   CSPROJ_PATH: src/Backend/AutopilotMonitor.Functions/AutopilotMonitor.Functions.csproj
 
 jobs:
-  build-and-deploy:
+  # Runs only this repository's scripts. The deploy identity trusts the `production`
+  # environment, which admits main only.
+  reserve:
+    name: Reserve build number
     runs-on: ubuntu-latest
-    name: Build and Deploy Backend
+    environment: production
+    permissions:
+      contents: read
+      id-token: write   # OIDC federated login (secret-less)
+    outputs:
+      build_number: ${{ steps.version.outputs.build_number }}
+      version: ${{ steps.version.outputs.version }}
+      build_utc: ${{ steps.version.outputs.build_utc }}
 
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
+      - name: Load endpoints
+        shell: pwsh
+        run: ./.github/scripts/Get-Endpoints.ps1 -ExportToGitHubEnv
+
       - name: Login to Azure (OIDC)
         uses: azure/login@a641126d1b8aa4d1fa005f4f92df94a3a4c4c906 # v3.1.0
         with:
@@ -30,11 +45,6 @@ jobs:
           tenant-id: ${{ secrets.AZURE_TENANT_ID }}
           subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}
 
-      - name: Setup .NET 10
-        uses: actions/setup-dotnet@a98b56852c35b8e3190ac28c8c2271da59106c68 # v6
-        with:
-          dotnet-version: '10.0.x'
-
       # Same counter mechanism as the agent line: reserve-before-build with ETag-CAS
       # against a per-component counter blob. BuildTimestampUtc is pinned here rather
       # than left to MSBuild's default so the published manifest and the value baked
@@ -43,7 +53,7 @@ jobs:
         id: version
         shell: pwsh
         run: |
-          $buildNumber = & ./.github/scripts/Request-BuildNumber.ps1 -CounterUrl "${{ env.VERSIONS_CONTAINER }}/backend.txt"
+          $buildNumber = & ./.github/scripts/Request-BuildNumber.ps1 -CounterUrl "$env:VERSIONS_CONTAINER_URL/backend.txt"
 
           $csproj = Get-Content ${{ env.CSPROJ_PATH }} -Raw
           if ($csproj -notmatch '<VersionPrefix>(\d+\.\d+)</VersionPrefix>') {
@@ -57,20 +67,90 @@ jobs:
           "version=$version" >> $env:GITHUB_OUTPUT
           "build_utc=$buildUtc" >> $env:GITHUB_OUTPUT
 
+  # Build zone: no login, no secret, no id-token. Its only output is the artifact; the
+  # deploy job treats that as data and never builds.
+  build:
+    name: Build
+    needs: reserve
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
+
+      - name: Setup .NET 10
+        uses: actions/setup-dotnet@a98b56852c35b8e3190ac28c8c2271da59106c68 # v6.0.0
+        with:
+          dotnet-version: '10.0.x'
+
       - name: Restore dependencies
         run: dotnet restore ${{ env.CSPROJ_PATH }}
 
       - name: Build
-        run: dotnet build ${{ env.CSPROJ_PATH }} --configuration Release --no-restore -p:SourceRevisionId=${{ github.sha }} -p:BuildNumber=${{ steps.version.outputs.build_number }} -p:BuildTimestampUtc=${{ steps.version.outputs.build_utc }}
+        run: dotnet build ${{ env.CSPROJ_PATH }} --configuration Release --no-restore -p:SourceRevisionId=${{ github.sha }} -p:BuildNumber=${{ needs.reserve.outputs.build_number }} -p:BuildTimestampUtc=${{ needs.reserve.outputs.build_utc }}
 
       - name: Publish
-        run: dotnet publish ${{ env.CSPROJ_PATH }} --configuration Release --output ./publish --no-build -p:SourceRevisionId=${{ github.sha }} -p:BuildNumber=${{ steps.version.outputs.build_number }} -p:BuildTimestampUtc=${{ steps.version.outputs.build_utc }}
+        run: dotnet publish ${{ env.CSPROJ_PATH }} --configuration Release --output ./publish --no-build -p:SourceRevisionId=${{ github.sha }} -p:BuildNumber=${{ needs.reserve.outputs.build_number }} -p:BuildTimestampUtc=${{ needs.reserve.outputs.build_utc }}
+
+      # One archive, not a folder upload: upload-artifact skips hidden files, and the
+      # isolated worker cannot start without the publish output's .azurefunctions/ folder.
+      - name: Package publish output
+        run: tar -czf "$RUNNER_TEMP/backend-publish.tgz" -C publish .
+
+      - name: Upload publish output
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
+        with:
+          name: backend-publish
+          path: ${{ runner.temp }}/backend-publish.tgz
+          if-no-files-found: error
+          retention-days: 1
+
+  # Trust zone: only this repository's scripts and pinned actions run here, never a restore
+  # or build. The package is unpacked outside the checkout, so nothing in it can replace the
+  # scripts this job runs.
+  deploy:
+    name: Deploy and verify
+    needs: [reserve, build]
+    runs-on: ubuntu-latest
+    environment: production
+    permissions:
+      contents: read
+      id-token: write   # OIDC federated login (secret-less)
+
+    steps:
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
+
+      - name: Load endpoints
+        shell: pwsh
+        run: ./.github/scripts/Get-Endpoints.ps1 -ExportToGitHubEnv
+
+      - name: Download publish output
+        uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
+        with:
+          name: backend-publish
+          path: ${{ runner.temp }}/backend-artifact
+
+      # GNU tar refuses members with '..' and strips leading '/'.
+      - name: Unpack publish output
+        run: |
+          mkdir "$RUNNER_TEMP/publish"
+          tar -xzf "$RUNNER_TEMP/backend-artifact/backend-publish.tgz" -C "$RUNNER_TEMP/publish"
+
+      - name: Login to Azure (OIDC)
+        uses: azure/login@a641126d1b8aa4d1fa005f4f92df94a3a4c4c906 # v3.1.0
+        with:
+          client-id: ${{ secrets.AZURE_CLIENT_ID }}
+          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
+          subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}
 
       - name: Deploy to Azure Functions
-        uses: Azure/functions-action@c5060b3b8bb1ebbcb531abd00c822ecbaa8ea656 # v1
+        uses: Azure/functions-action@c5060b3b8bb1ebbcb531abd00c822ecbaa8ea656 # v1.5.7
         with:
           app-name: ${{ env.FUNCTION_APP }}
-          package: ./publish
+          package: ${{ runner.temp }}/publish
 
       # A green deploy step only means the package was accepted. This gate makes the
       # workflow assert what is actually being served, so the manifest published below
@@ -78,16 +158,18 @@ jobs:
       # /api/health is IP rate-limited at 30/min; one poll per 10s stays well under.
       - name: Verify deployed build is live
         shell: pwsh
+        env:
+          EXPECTED_VERSION: ${{ needs.reserve.outputs.version }}
         run: |
           $expectedCommit = '${{ github.sha }}'.Substring(0, 7)
-          $expectedVersion = '${{ steps.version.outputs.version }}'
+          $expectedVersion = $env:EXPECTED_VERSION
           $deadline = [DateTime]::UtcNow.AddMinutes(3)
           $attempt = 0
 
           while ([DateTime]::UtcNow -lt $deadline) {
             $attempt++
             try {
-              $health = Invoke-RestMethod -Uri '${{ env.HEALTH_URL }}' -Method Get -TimeoutSec 30
+              $health = Invoke-RestMethod -Uri "$env:API_BASE_URL/api/health" -Method Get -TimeoutSec 30
               if ($health.commitHash -eq $expectedCommit -and $health.version -eq $expectedVersion) {
                 Write-Host "Live after $attempt attempt(s): $($health.version) @ $($health.commitHash), built $($health.buildUtc)"
                 exit 0
@@ -107,18 +189,21 @@ jobs:
       # (rollback, stale instance) rather than a half-finished pipeline.
       - name: Publish version manifest
         shell: pwsh
+        env:
+          BACKEND_VERSION: ${{ needs.reserve.outputs.version }}
+          BACKEND_BUILD_UTC: ${{ needs.reserve.outputs.build_utc }}
         run: |
           $manifest = [ordered]@{
             component   = 'backend'
-            version     = '${{ steps.version.outputs.version }}'
+            version     = $env:BACKEND_VERSION
             commit      = '${{ github.sha }}'.Substring(0, 7)
-            buildUtc    = '${{ steps.version.outputs.build_utc }}'
+            buildUtc    = $env:BACKEND_BUILD_UTC
             deployedUtc = [DateTime]::UtcNow.ToString('o')
             runId       = '${{ github.run_id }}'
           } | ConvertTo-Json
 
           $authHeaders = & ./.github/scripts/Get-StorageAuthHeaders.ps1
-          $url = "${{ env.VERSIONS_CONTAINER }}/backend.json"
+          $url = "$env:VERSIONS_CONTAINER_URL/backend.json"
           # no-cache: rotating pointer, must never be served stale by any fronting cache.
           Invoke-RestMethod -Uri $url -Method Put -Headers ($authHeaders + @{ 'x-ms-blob-type' = 'BlockBlob'; 'x-ms-blob-cache-control' = 'no-cache' }) -Body $manifest -ContentType 'application/json' | Out-Null
           Write-Host "Published versions/backend.json"

.github/workflows/deploy-mcp.yml

@@ -16,20 +16,29 @@ env:
   # reference — with a sha-only tag `az containerapp update` saw the same image
   # and silently created no revision (2026-08-28).
   IMAGE_TAG: ${{ github.sha }}-${{ github.run_number }}
-  HEALTH_URL: https://mcp.autopilotmonitor.com/health
-  VERSIONS_CONTAINER: https://autopilotmonitoreu.blob.core.windows.net/versions
   PACKAGE_JSON: src/McpServer/autopilot-monitor-mcp/package.json
+  # Hosts (MCP, versions container) come from .github/scripts/Get-Endpoints.ps1.
 
 jobs:
+  # One job on purpose, unlike the other deploys: all third-party code (npm ci, the build, the
+  # embedding precompute) runs inside `docker build`, whose containers see neither the az
+  # login nor the registry credentials nor this job's OIDC request token. That container is
+  # the boundary -- never move npm onto the runner here, or the build gets the job's
+  # credentials. The deploy identity trusts the `production` environment, which admits main only.
   build-and-deploy:
     runs-on: ubuntu-latest
     name: Build and Deploy MCP Server
+    environment: production
 
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
+      - name: Load endpoints
+        shell: pwsh
+        run: ./.github/scripts/Get-Endpoints.ps1 -ExportToGitHubEnv
+
       # The published customer documentation lives in its own PUBLIC repo and is
       # baked into the image as the search_docs corpus (see the Dockerfile). No
       # token is needed. Checking it out here rather than vendoring it keeps the
@@ -38,7 +47,7 @@ jobs:
       # re-run actually rolls a new revision). /health reports the SHA captured
       # below so that staleness is checkable.
       - name: Check out documentation bundle
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
           repository: autopilotmonitor/autopilotmonitor-docs
@@ -62,7 +71,7 @@ jobs:
         id: version
         shell: pwsh
         run: |
-          $buildNumber = & ./.github/scripts/Request-BuildNumber.ps1 -CounterUrl "${{ env.VERSIONS_CONTAINER }}/mcp.txt"
+          $buildNumber = & ./.github/scripts/Request-BuildNumber.ps1 -CounterUrl "$env:VERSIONS_CONTAINER_URL/mcp.txt"
 
           $pkgVersion = (Get-Content ${{ env.PACKAGE_JSON }} -Raw | ConvertFrom-Json).version
           $prefix = ($pkgVersion -split '\.')[0..1] -join '.'
@@ -117,7 +126,7 @@ jobs:
           while ([DateTime]::UtcNow -lt $deadline) {
             $attempt++
             try {
-              $health = Invoke-RestMethod -Uri '${{ env.HEALTH_URL }}' -Method Get -TimeoutSec 60
+              $health = Invoke-RestMethod -Uri "$env:MCP_BASE_URL/health" -Method Get -TimeoutSec 60
               if ($health.commit -eq $expectedCommit -and $health.version -eq $expectedVersion) {
                 Write-Host "Live after $attempt attempt(s): $($health.version) @ $($health.commit), docs $($health.docs.commit) ($($health.docs.chunks) chunks)"
                 exit 0
@@ -148,7 +157,7 @@ jobs:
           } | ConvertTo-Json
 
           $authHeaders = & ./.github/scripts/Get-StorageAuthHeaders.ps1
-          $url = "${{ env.VERSIONS_CONTAINER }}/mcp.json"
+          $url = "$env:VERSIONS_CONTAINER_URL/mcp.json"
           # no-cache: rotating pointer, must never be served stale by any fronting cache.
           Invoke-RestMethod -Uri $url -Method Put -Headers ($authHeaders + @{ 'x-ms-blob-type' = 'BlockBlob'; 'x-ms-blob-cache-control' = 'no-cache' }) -Body $manifest -ContentType 'application/json' | Out-Null
           Write-Host "Published versions/mcp.json"

.github/workflows/deploy-web.yml

@@ -5,6 +5,11 @@ name: Deploy Web (Static Web App)
 # path ci.yml tests (setup-node 22 + npm ci + next build) instead of Oryx guessing inside the
 # action's container, and the exported site is attested like the agent ZIP and the MSI.
 #
+# Three jobs, each holding only what it needs: build runs all third-party code without any
+# credential, deploy holds the SWA token alone, record attests, verifies and writes the
+# manifest. A failed record job is re-run on its own; it writes only once the live site
+# serves this commit.
+#
 # PR preview environments were dropped with the action: their per-PR URLs are never
 # registered as redirect URIs, so nobody could log in on them anyway.
 
@@ -17,38 +22,39 @@ on:
       - '.github/workflows/deploy-web.yml'
   workflow_dispatch:
 
+# Every job states its own permissions. This default is what the build job runs with: npm ci
+# (install scripts included) and next build execute third-party code and must never hold more
+# than a read-only token.
 permissions:
   contents: read
-  id-token: write      # keyless Sigstore signing (artifact attestation)
-  attestations: write  # store the provenance attestation
 
+# Hosts (API, portal, storage) come from .github/scripts/Get-Endpoints.ps1, never from a
+# repository variable.
 env:
-  PORTAL_VERSION_URL: https://portal.autopilotmonitor.com/version.json
-  VERSIONS_CONTAINER: https://autopilotmonitoreu.blob.core.windows.net/versions
   WEB_DIR: src/Web/autopilot-monitor-web
   # Exact version: `npx --package` resolves from the registry at run time, so the pin is the
   # only thing that keeps the deploy step's code fixed. Bump deliberately.
   SWA_CLI_VERSION: 2.0.10
 
 jobs:
-  build_and_deploy_job:
+  # Build zone: no login, no id-token. The one secret here is the Application Insights
+  # connection string, which ships inside the public bundle anyway.
+  build:
+    name: Build
     runs-on: ubuntu-latest
-    name: Build and Deploy Job
+    outputs:
+      docs_sha: ${{ steps.docs.outputs.sha }}
+
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
           lfs: false
           fetch-depth: 0  # Full history needed for git-based sitemap lastmod dates
 
-      # Only the version manifest at the very end needs Azure; logging in here rather than
-      # next to that step keeps a credential problem from surfacing after a live deploy.
-      - name: Login to Azure (OIDC)
-        uses: azure/login@a641126d1b8aa4d1fa005f4f92df94a3a4c4c906 # v3.1.0
-        with:
-          client-id: ${{ secrets.AZURE_CLIENT_ID }}
-          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
-          subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}
+      - name: Load endpoints
+        shell: pwsh
+        run: ./.github/scripts/Get-Endpoints.ps1 -ExportToGitHubEnv
 
       # The customer docs repo is the single source of the "What's new" panel: the prebuild
       # step (scripts/whats-new/build-whats-new.js) turns changelog/*.md into
@@ -57,7 +63,7 @@ jobs:
       # workflow runs again (push under src/Web/** or a manual dispatch); versions/web.json
       # records which docs commit is live.
       - name: Check out documentation bundle
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
           repository: autopilotmonitor/autopilotmonitor-docs
@@ -68,7 +74,7 @@ jobs:
         id: docs
         run: echo "sha=$(git -C whats-new-docs rev-parse HEAD)" >> "$GITHUB_OUTPUT"
 
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
           node-version: '22'
           cache: npm
@@ -105,12 +111,6 @@ jobs:
           # and the default app for browsers without a stored selection (legacy|primary).
           NEXT_PUBLIC_ENTRA_LEGACY_CLIENT_ID: ${{ vars.NEXT_PUBLIC_ENTRA_LEGACY_CLIENT_ID }}
           NEXT_PUBLIC_ENTRA_DEFAULT_APP: ${{ vars.NEXT_PUBLIC_ENTRA_DEFAULT_APP }}
-          NEXT_PUBLIC_ENTRA_REDIRECT_URI: ${{ vars.NEXT_PUBLIC_ENTRA_REDIRECT_URI }}
-          NEXT_PUBLIC_ENTRA_POST_LOGOUT_REDIRECT_URI: ${{ vars.NEXT_PUBLIC_ENTRA_POST_LOGOUT_REDIRECT_URI }}
-          NEXT_PUBLIC_API_BASE_URL: ${{ vars.NEXT_PUBLIC_API_BASE_URL }}
-          # Browsers load the live stats through it, and the build reads it once to bake the
-          # numbers of this moment into the landing page's static HTML (lib/platformStats.ts).
-          NEXT_PUBLIC_PLATFORM_STATS_MANIFEST_URL: ${{ vars.NEXT_PUBLIC_PLATFORM_STATS_MANIFEST_URL }}
           NEXT_PUBLIC_APPINSIGHTS_CONNECTION_STRING: ${{ secrets.NEXT_PUBLIC_APPINSIGHTS_CONNECTION_STRING }}
           # Build stamp for public/version.json (scripts/generate-version.js in prebuild);
           # the verify step below polls the deployed file for exactly this commit.
@@ -120,6 +120,18 @@ jobs:
           # Star counter from the step above; empty when the lookup failed.
           NEXT_PUBLIC_GITHUB_STARS: ${{ steps.stars.outputs.count }}
         run: |
+          # All from the endpoint registry (Load endpoints). Empty, the portal would silently call
+          # localhost:7071. The stats manifest is what browsers load the live numbers from, and the
+          # build reads it once to bake the numbers of this moment into the landing page's static
+          # HTML (lib/platformStats.ts). The redirect URIs are only the fallback for the static
+          # render; in the browser lib/msalConfig.ts uses window.location.origin.
+          export NEXT_PUBLIC_API_BASE_URL="$API_BASE_URL"
+          export NEXT_PUBLIC_PLATFORM_STATS_MANIFEST_URL="$STATS_CONTAINER_URL"
+          export NEXT_PUBLIC_ENTRA_REDIRECT_URI="$WWW_BASE_URL"
+          export NEXT_PUBLIC_ENTRA_POST_LOGOUT_REDIRECT_URI="$WWW_BASE_URL"
+          test -n "$NEXT_PUBLIC_API_BASE_URL" || { echo "::error::API_BASE_URL is empty - the portal would call localhost"; exit 1; }
+          test -n "$NEXT_PUBLIC_PLATFORM_STATS_MANIFEST_URL" || { echo "::error::STATS_CONTAINER_URL is empty"; exit 1; }
+          test -n "$NEXT_PUBLIC_ENTRA_REDIRECT_URI" || { echo "::error::WWW_BASE_URL is empty"; exit 1; }
           npm run build
           test -f out/version.json || { echo "out/version.json missing - export did not run"; exit 1; }
           test -f out/whats-new.json || { echo "out/whats-new.json missing - whats-new prebuild did not run"; exit 1; }
@@ -136,19 +148,37 @@ jobs:
           echo "archive=$archive" >> "$GITHUB_OUTPUT"
           sha256sum "$archive"
 
-      # Sigstore keyless provenance: binds the export digest to this workflow run (repo, ref,
-      # commit, workflow file) in a public transparency log.
-      # Verify: gh attestation verify portal-<sha>.tar.gz --repo okieselbach/AutopilotMonitor
-      - name: Attest build provenance (portal export)
-        uses: actions/attest-build-provenance@4d101475d8b20a2381f78447822ac1eab6504dd8 # v4
-        with:
-          subject-path: ${{ steps.package.outputs.archive }}
-
       - name: Upload build artifact
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
         with:
           name: portal-${{ github.sha }}
           path: ${{ steps.package.outputs.archive }}
+          if-no-files-found: error
+
+  # Holds the SWA deployment token and nothing else -- deliberately no id-token: the SWA CLI
+  # is third-party npm code whose dependencies resolve at run time.
+  deploy:
+    name: Deploy to Static Web App
+    needs: build
+    runs-on: ubuntu-latest
+    environment: production
+
+    steps:
+      # For staticwebapp.config.json only.
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
+
+      - name: Download build artifact
+        uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
+        with:
+          name: portal-${{ github.sha }}
+          path: ${{ runner.temp }}/portal
+
+      - name: Unpack export
+        run: |
+          mkdir "$WEB_DIR/out"
+          tar -xzf "$RUNNER_TEMP/portal/portal-${GITHUB_SHA::7}.tar.gz" -C "$WEB_DIR/out"
 
       # Upload only -- out/ is already built. --swa-config-location points at the app root,
       # where staticwebapp.config.json lives (next export does not copy it into out/).
@@ -163,6 +193,48 @@ jobs:
             --env production \
             --no-use-keychain
 
+  # Trust zone: only this repository's scripts and pinned actions run here. The deploy
+  # identity trusts the `production` environment, which admits main only.
+  record:
+    name: Attest, verify and record
+    needs: [build, deploy]
+    runs-on: ubuntu-latest
+    environment: production
+    permissions:
+      contents: read
+      id-token: write      # keyless Sigstore signing, OIDC login for the manifest
+      attestations: write  # store the provenance attestation
+
+    steps:
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
+
+      - name: Load endpoints
+        shell: pwsh
+        run: ./.github/scripts/Get-Endpoints.ps1 -ExportToGitHubEnv
+
+      - name: Download build artifact
+        uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
+        with:
+          name: portal-${{ github.sha }}
+          path: ${{ runner.temp }}/portal
+
+      - name: Locate the export archive
+        id: archive
+        run: |
+          path="$RUNNER_TEMP/portal/portal-${GITHUB_SHA::7}.tar.gz"
+          test -f "$path" || { echo "::error::Export archive for this commit is missing"; exit 1; }
+          echo "path=$path" >> "$GITHUB_OUTPUT"
+
+      # Sigstore keyless provenance: binds the export digest to this workflow run (repo, ref,
+      # commit, workflow file) in a public transparency log.
+      # Verify: gh attestation verify portal-<sha>.tar.gz --repo okieselbach/AutopilotMonitor
+      - name: Attest build provenance (portal export)
+        uses: actions/attest-build-provenance@4d101475d8b20a2381f78447822ac1eab6504dd8 # v4.2.2
+        with:
+          subject-path: ${{ steps.archive.outputs.path }}
+
       # The CLI reporting success proves the upload, not that the production host serves the
       # new build -- verify against the live /version.json (Cache-Control: no-store via
       # staticwebapp.config.json) before publishing the version manifest.
@@ -176,7 +248,7 @@ jobs:
           while ([DateTime]::UtcNow -lt $deadline) {
             $attempt++
             try {
-              $v = Invoke-RestMethod -Uri '${{ env.PORTAL_VERSION_URL }}' -Method Get -TimeoutSec 30 -Headers @{ 'Cache-Control' = 'no-cache' }
+              $v = Invoke-RestMethod -Uri "$env:PORTAL_BASE_URL/version.json" -Method Get -TimeoutSec 30 -Headers @{ 'Cache-Control' = 'no-cache' }
               if ($v.commit -eq $expectedCommit) {
                 Write-Host "Live after $attempt attempt(s): web @ $($v.commit), built $($v.buildUtc)"
                 exit 0
@@ -190,22 +262,36 @@ jobs:
 
           throw "Portal did not serve commit $expectedCommit within 5 minutes -- deployment not verified, version manifest NOT published"
 
+      - name: Login to Azure (OIDC)
+        uses: azure/login@a641126d1b8aa4d1fa005f4f92df94a3a4c4c906 # v3.1.0
+        with:
+          client-id: ${{ secrets.AZURE_CLIENT_ID }}
+          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
+          subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}
+
       # Cross-component deploy record, written last and only on a verified deploy --
       # same contract as versions/backend.json and versions/mcp.json. docsCommit says which
       # docs revision the live What's new panel was generated from (as for the MCP corpus).
+      # The docs commit comes from the build job, so it arrives as data and is checked
+      # against the shape of a commit id before it is used.
       - name: Publish version manifest
         shell: pwsh
+        env:
+          DOCS_SHA: ${{ needs.build.outputs.docs_sha }}
         run: |
+          if ($env:DOCS_SHA -notmatch '^[0-9a-f]{40}$') {
+            throw "The build job handed over no valid docs commit -- version manifest NOT published."
+          }
           $manifest = [ordered]@{
             component   = 'web'
             commit      = '${{ github.sha }}'.Substring(0, 7)
             deployedUtc = [DateTime]::UtcNow.ToString('o')
             runId       = '${{ github.run_id }}'
-            docsCommit  = '${{ steps.docs.outputs.sha }}'.Substring(0, 7)
+            docsCommit  = $env:DOCS_SHA.Substring(0, 7)
           } | ConvertTo-Json
 
           $authHeaders = & ./.github/scripts/Get-StorageAuthHeaders.ps1
-          $url = "${{ env.VERSIONS_CONTAINER }}/web.json"
+          $url = "$env:VERSIONS_CONTAINER_URL/web.json"
           # no-cache: rotating pointer, must never be served stale by any fronting cache.
           Invoke-RestMethod -Uri $url -Method Put -Headers ($authHeaders + @{ 'x-ms-blob-type' = 'BlockBlob'; 'x-ms-blob-cache-control' = 'no-cache' }) -Body $manifest -ContentType 'application/json' | Out-Null
           Write-Host "Published versions/web.json"

.github/workflows/indexnow-notify.yml

@@ -3,24 +3,34 @@ name: IndexNow Notify (Bing/Yandex)
 on:
   workflow_dispatch:
 
-# The job only talks to public endpoints; it never needs a repository token.
-permissions: {}
+# Read-only, for the checkout of the endpoint registry; beyond that the job only talks to
+# public endpoints.
+permissions:
+  contents: read
 
 jobs:
   notify:
     runs-on: ubuntu-latest
     name: Submit URLs to IndexNow
     steps:
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
+
+      - name: Load endpoints
+        shell: pwsh
+        run: ./.github/scripts/Get-Endpoints.ps1 -ExportToGitHubEnv
+
       - name: Notify IndexNow
         run: |
-          URLS=$(curl -s https://www.autopilotmonitor.com/sitemap.xml | grep -oP '<loc>\K[^<]+')
+          URLS=$(curl -s "$WWW_BASE_URL/sitemap.xml" | grep -oP '<loc>\K[^<]+')
           URL_JSON=$(echo "$URLS" | jq -R -s 'split("\n") | map(select(length > 0))')
           RESPONSE=$(curl -s -o /dev/null -w "%{http_code}" -X POST "https://api.indexnow.org/indexnow" \
             -H "Content-Type: application/json" \
             -d "{
-              \"host\": \"www.autopilotmonitor.com\",
+              \"host\": \"${WWW_BASE_URL#https://}\",
               \"key\": \"8f1710b5-f14b-477e-938b-6c01ef851661\",
-              \"keyLocation\": \"https://www.autopilotmonitor.com/IndexNow.txt\",
+              \"keyLocation\": \"$WWW_BASE_URL/IndexNow.txt\",
               \"urlList\": $URL_JSON
             }")
           echo "IndexNow response: HTTP $RESPONSE"

.github/workflows/oidc-login-test.yml

@@ -1,28 +0,0 @@
-# TEMPORARY: proves the secret-less OIDC federated login works end-to-end.
-# Delete after the cross-tenant migration is verified (or keep as a cheap smoke test).
-name: OIDC Login Test
-
-on:
-  workflow_dispatch:
-
-permissions:
-  id-token: write
-  contents: read
-
-jobs:
-  login-test:
-    runs-on: ubuntu-latest
-    name: Federated login + read check
-    steps:
-      - name: Login to Azure (OIDC)
-        uses: azure/login@a641126d1b8aa4d1fa005f4f92df94a3a4c4c906 # v3.1.0
-        with:
-          client-id: ${{ secrets.AZURE_CLIENT_ID }}
-          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
-          subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}
-
-      - name: Prove identity and RBAC scope
-        run: |
-          az account show --query "{subscription:name, subscriptionId:id, tenantId:tenantId, user:user.name}" -o json
-          echo "--- function app visible with Contributor scope? ---"
-          az functionapp list --query "[].name" -o json

.github/workflows/publish-scripts.yml

@@ -36,13 +36,13 @@ concurrency:
 
 # Authenticode for the customer scripts. Same identity and same repository variables as
 # build-agent.yml -- the loader on every device pins this publisher, so the two must not
-# drift apart.
+# drift apart. The publisher takes every host, the timestamp server included, from
+# .github/scripts/Get-Endpoints.ps1.
 env:
   CODE_SIGN_CLIENT_ID: ${{ vars.CODE_SIGN_CLIENT_ID }}
   CODE_SIGN_CERTIFICATE_NAME: ${{ vars.CODE_SIGN_CERTIFICATE_NAME }}
   CODE_SIGN_KEY_VAULT_URL: ${{ vars.CODE_SIGN_KEY_VAULT_URL }}
   CODE_SIGN_EXPECTED_SUBJECT: ${{ vars.CODE_SIGN_EXPECTED_SUBJECT }}
-  CODE_SIGN_TIMESTAMP_URL: http://timestamp.acs.microsoft.com
 
 jobs:
   # $/ is the self-repository form: this repository at the running commit, never a
@@ -60,7 +60,7 @@ jobs:
       contents: read
       id-token: write   # OIDC federated login for the storage writes (secret-less)
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
 
@@ -124,7 +124,6 @@ jobs:
             -SigningToken $env:CODE_SIGN_AKV_ACCESS_TOKEN `
             -KeyVaultUrl $env:CODE_SIGN_KEY_VAULT_URL `
             -CertificateName $env:CODE_SIGN_CERTIFICATE_NAME `
-            -TimestampUrl $env:CODE_SIGN_TIMESTAMP_URL `
             -ExpectedPublisher $env:CODE_SIGN_EXPECTED_SUBJECT `
             -SignedOutputDir "$env:RUNNER_TEMP/signed-scripts" `
             -DryRun:$dryRun
@@ -135,7 +134,7 @@ jobs:
       # when someone wants to look at the file.
       - name: Upload signed scripts
         if: always()
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
         with:
           name: signed-scripts
           path: ${{ runner.temp }}/signed-scripts

AutopilotMonitor.sln

@@ -31,6 +31,12 @@ Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "AutopilotMonitor.Agent.V2.C
 EndProject
 Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "AutopilotMonitor.SummaryDialog.Tests", "src\Agent\AutopilotMonitor.SummaryDialog.Tests\AutopilotMonitor.SummaryDialog.Tests.csproj", "{4F1234C1-9387-4D2C-93E0-713698A2DFCB}"
 EndProject
+Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Push", "Push", "{09266D87-DDB8-2F61-A22F-10D1DD295F4F}"
+EndProject
+Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "AutopilotMonitor.Push", "src\Push\AutopilotMonitor.Push\AutopilotMonitor.Push.csproj", "{2C3E71E4-0049-45A5-839D-F4171B47D5F3}"
+EndProject
+Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "AutopilotMonitor.Push.Tests", "src\Push\AutopilotMonitor.Push.Tests\AutopilotMonitor.Push.Tests.csproj", "{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}"
+EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
 		Debug|Any CPU = Debug|Any CPU
@@ -161,6 +167,30 @@ Global
 		{4F1234C1-9387-4D2C-93E0-713698A2DFCB}.Release|x64.Build.0 = Release|Any CPU
 		{4F1234C1-9387-4D2C-93E0-713698A2DFCB}.Release|x86.ActiveCfg = Release|Any CPU
 		{4F1234C1-9387-4D2C-93E0-713698A2DFCB}.Release|x86.Build.0 = Release|Any CPU
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3}.Debug|Any CPU.ActiveCfg = Debug|Any CPU
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3}.Debug|Any CPU.Build.0 = Debug|Any CPU
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3}.Debug|x64.ActiveCfg = Debug|Any CPU
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3}.Debug|x64.Build.0 = Debug|Any CPU
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3}.Debug|x86.ActiveCfg = Debug|Any CPU
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3}.Debug|x86.Build.0 = Debug|Any CPU
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3}.Release|Any CPU.ActiveCfg = Release|Any CPU
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3}.Release|Any CPU.Build.0 = Release|Any CPU
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3}.Release|x64.ActiveCfg = Release|Any CPU
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3}.Release|x64.Build.0 = Release|Any CPU
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3}.Release|x86.ActiveCfg = Release|Any CPU
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3}.Release|x86.Build.0 = Release|Any CPU
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}.Debug|Any CPU.ActiveCfg = Debug|Any CPU
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}.Debug|Any CPU.Build.0 = Debug|Any CPU
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}.Debug|x64.ActiveCfg = Debug|Any CPU
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}.Debug|x64.Build.0 = Debug|Any CPU
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}.Debug|x86.ActiveCfg = Debug|Any CPU
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}.Debug|x86.Build.0 = Debug|Any CPU
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}.Release|Any CPU.ActiveCfg = Release|Any CPU
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}.Release|Any CPU.Build.0 = Release|Any CPU
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}.Release|x64.ActiveCfg = Release|Any CPU
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}.Release|x64.Build.0 = Release|Any CPU
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}.Release|x86.ActiveCfg = Release|Any CPU
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302}.Release|x86.Build.0 = Release|Any CPU
 	EndGlobalSection
 	GlobalSection(SolutionProperties) = preSolution
 		HideSolutionNode = FALSE
@@ -179,5 +209,8 @@ Global
 		{11B36E5C-B512-4D1E-A76B-E5AD866B033C} = {096B6098-F141-7114-B090-72D1899D8853}
 		{95A09BF0-4E22-41D9-9B43-8D6ECC2C8AD0} = {096B6098-F141-7114-B090-72D1899D8853}
 		{4F1234C1-9387-4D2C-93E0-713698A2DFCB} = {096B6098-F141-7114-B090-72D1899D8853}
+		{09266D87-DDB8-2F61-A22F-10D1DD295F4F} = {827E0CD3-B72D-47B6-A68D-7590B98EB39B}
+		{2C3E71E4-0049-45A5-839D-F4171B47D5F3} = {09266D87-DDB8-2F61-A22F-10D1DD295F4F}
+		{6BB95C52-F9D5-4C48-A107-B2B1FD5EE302} = {09266D87-DDB8-2F61-A22F-10D1DD295F4F}
 	EndGlobalSection
 EndGlobal

Directory.Build.props

@@ -8,5 +8,8 @@
     <TreatWarningsAsErrors>true</TreatWarningsAsErrors>
     <MSBuildTreatWarningsAsErrors>true</MSBuildTreatWarningsAsErrors>
     <WarningsNotAsErrors>NU1901;NU1902;NU1903;NU1904</WarningsNotAsErrors>
+    <!-- Run the IDE code-style analyzers in the build too, so the rules .editorconfig raises
+         to warning fail it like any compiler warning. -->
+    <EnforceCodeStyleInBuild>true</EnforceCodeStyleInBuild>
   </PropertyGroup>
 </Project>

rules/analyze/ANALYZE-APP-001.json

@@ -5,7 +5,7 @@
   "description": "Detects when a Win32 app detection script fails or returns unexpected results.",
   "severity": "warning",
   "category": "apps",
-  "version": "1.2.0",
+  "version": "1.3.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -38,6 +38,26 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "{{detection_failure.appName}} was not detected after its install",
+    "subject": {
+      "kind": "app",
+      "key": "{{detection_failure.appId}}",
+      "name": "{{detection_failure.appName}}"
+    },
+    "facts": [
+      {
+        "label": "Detection",
+        "value": "{{detection_failure.value}}"
+      },
+      {
+        "label": "Error",
+        "value": "{{detection_failure.errorCode}}",
+        "format": "errorCode"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "detection",

rules/analyze/ANALYZE-APP-002.json

@@ -5,7 +5,7 @@
   "description": "Detects app installation failures with specific exit codes or HRESULT error codes and suggests fixes.",
   "severity": "high",
   "category": "apps",
-  "version": "1.2.0",
+  "version": "1.3.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -41,6 +41,27 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "{{install_error.appName}} failed to install",
+    "subject": {
+      "kind": "app",
+      "key": "{{install_error.appId}}",
+      "name": "{{install_error.appName}}"
+    },
+    "facts": [
+      {
+        "label": "Exit code",
+        "value": "{{install_error.exitCode}}",
+        "format": "code"
+      },
+      {
+        "label": "Error",
+        "value": "{{install_error.errorCode}}",
+        "format": "errorCode"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "msi",

rules/analyze/ANALYZE-APP-003.json

@@ -33,9 +33,25 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "{{dependency_failure.appName}} failed because a dependency did not install",
+    "subject": {
+      "kind": "app",
+      "key": "{{dependency_failure.appId}}",
+      "name": "{{dependency_failure.appName}}"
+    },
+    "facts": [
+      {
+        "label": "Error",
+        "value": "{{dependency_failure.errorCode}}",
+        "format": "errorCode"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "dependency"
   ],
-  "version": "1.1.0"
+  "version": "1.2.0"
 }

rules/analyze/ANALYZE-APP-004.json

@@ -5,7 +5,7 @@
   "description": "Detects when disk space is critically low during enrollment, which can cause app installation failures. Confidence increases when app installations actually failed while disk space was low.",
   "severity": "critical",
   "category": "apps",
-  "version": "1.2.0",
+  "version": "1.3.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -71,6 +71,27 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "Free disk space dropped to {{low_disk.value}} GB",
+    "subject": {
+      "kind": "device",
+      "key": "disk",
+      "name": "Disk space"
+    },
+    "facts": [
+      {
+        "label": "Free space (GB)",
+        "value": "{{low_disk.value}}",
+        "format": "number"
+      },
+      {
+        "label": "Measured",
+        "value": "{{low_disk.timestamp}}",
+        "format": "timestamp"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "disk-space",

rules/analyze/ANALYZE-APP-005.json

@@ -33,10 +33,26 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "The content download of {{download_timeout.appName}} timed out or failed",
+    "subject": {
+      "kind": "app",
+      "key": "{{download_timeout.appId}}",
+      "name": "{{download_timeout.appName}}"
+    },
+    "facts": [
+      {
+        "label": "Error",
+        "value": "{{download_timeout.errorCode}}",
+        "format": "errorCode"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "download",
     "timeout"
   ],
-  "version": "1.1.0"
+  "version": "1.2.0"
 }

rules/analyze/ANALYZE-APP-006.json

@@ -5,7 +5,7 @@
   "description": "Detects when the same app installation is attempted 3 or more times during one enrollment — a retry loop caused by unhonored reboot requests, repeated failures, or detection mismatches.",
   "severity": "high",
   "category": "apps",
-  "version": "1.2.0",
+  "version": "1.3.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -35,6 +35,27 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "{{install_retry_loop.appName}} was started {{install_retry_loop.count}} times",
+    "subject": {
+      "kind": "app",
+      "key": "{{install_retry_loop.groupKey}}",
+      "name": "{{install_retry_loop.appName}}"
+    },
+    "facts": [
+      {
+        "label": "Install starts",
+        "value": "{{install_retry_loop.count}}",
+        "format": "number"
+      },
+      {
+        "label": "First start",
+        "value": "{{install_retry_loop.timestamp}}",
+        "format": "timestamp"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "reboot",

rules/analyze/ANALYZE-APP-007.json

@@ -5,7 +5,7 @@
   "description": "Detects when the app tracking summary reports multiple failed app installations.",
   "severity": "high",
   "category": "apps",
-  "version": "1.2.0",
+  "version": "1.3.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -50,6 +50,17 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "{{summary_errors.value}} app installations failed",
+    "facts": [
+      {
+        "label": "Failed apps",
+        "value": "{{summary_errors.value}}",
+        "format": "number"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "summary",

rules/analyze/ANALYZE-APP-008.json

@@ -32,11 +32,32 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "{{slow_app_install.appName}} took {{slow_app_install.durationFormatted}} to install",
+    "subject": {
+      "kind": "app",
+      "key": "{{slow_app_install.appId}}",
+      "name": "{{slow_app_install.appName}}"
+    },
+    "facts": [
+      {
+        "label": "Install time",
+        "value": "{{slow_app_install.durationSeconds}}",
+        "format": "duration"
+      },
+      {
+        "label": "Started",
+        "value": "{{slow_app_install.startTimestamp}}",
+        "format": "timestamp"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "performance",
     "slow-install",
     "warning"
   ],
-  "version": "1.1.0"
+  "version": "1.2.0"
 }

rules/analyze/ANALYZE-APP-010.json

@@ -1,6 +1,7 @@
 {
   "$schema": "../schema/analyze-rule.schema.json",
   "ruleId": "ANALYZE-APP-010",
+  "version": "1.1.0",
   "title": "App Reverted to Error After Successful Install",
   "description": "Detects when a specific app (matched by appId) was first reported as successfully installed but subsequently reported an IME enforcement error — while enrollment completed successfully. The app may be broken post-enrollment.",
   "severity": "warning",
@@ -56,5 +57,31 @@
     { "title": "Troubleshoot Win32 app installations", "url": "https://learn.microsoft.com/en-us/troubleshoot/mem/intune/app-management/troubleshoot-win32-app-install" },
     { "title": "Intune Management Extension logs", "url": "https://learn.microsoft.com/en-us/mem/intune/apps/intune-management-extension" }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "{{app_regression.eventA_appName}} installed, then reported an error",
+    "subject": {
+      "kind": "app",
+      "key": "{{app_regression.joinValue}}",
+      "name": "{{app_regression.eventA_appName}}"
+    },
+    "facts": [
+      {
+        "label": "Installed",
+        "value": "{{app_regression.eventA_timestamp}}",
+        "format": "timestamp"
+      },
+      {
+        "label": "Error after",
+        "value": "{{app_regression.timeDeltaSeconds}}",
+        "format": "duration"
+      },
+      {
+        "label": "Error",
+        "value": "{{app_regression.eventB_errorCode}}",
+        "format": "errorCode"
+      }
+    ]
+  },
   "tags": ["apps", "win32", "enforcement", "post-install", "correlation"]
 }

rules/analyze/ANALYZE-APP-012.json

@@ -43,6 +43,32 @@
       "url": "https://learn.microsoft.com/en-us/mem/intune/apps/apps-win32-app-management#step-4-return-codes"
     }
   ],
+  "presentation": {
+    "role": "advisory",
+    "summary": "{{enforcement_error_resolved.eventA_appName}} reported an install error but was detected as installed",
+    "subject": {
+      "kind": "app",
+      "key": "{{enforcement_error_resolved.joinValue}}",
+      "name": "{{enforcement_error_resolved.eventA_appName}}"
+    },
+    "facts": [
+      {
+        "label": "Exit code",
+        "value": "{{enforcement_error_resolved.eventA_exitCode}}",
+        "format": "code"
+      },
+      {
+        "label": "Error",
+        "value": "{{enforcement_error_resolved.eventA_errorCode}}",
+        "format": "errorCode"
+      },
+      {
+        "label": "Detected installed after",
+        "value": "{{enforcement_error_resolved.timeDeltaSeconds}}",
+        "format": "duration"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "enforcement",
@@ -50,5 +76,5 @@
     "unmapped",
     "resolved"
   ],
-  "version": "1.1.0"
+  "version": "1.2.0"
 }

rules/analyze/ANALYZE-APP-013.json

@@ -5,7 +5,7 @@
   "description": "Detects when the Enrollment Status Page failed because Intune could not verify an app was installed (HRESULT 0x87D1041C). The installer ran but the detection rule did not match afterwards.",
   "severity": "critical",
   "category": "apps",
-  "version": "1.2.0",
+  "version": "1.3.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -82,6 +82,26 @@
       "url": "https://learn.microsoft.com/en-us/mem/intune/apps/apps-win32-add#step-4-detection-rules"
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "{{app_with_detection_failure.appName}} installed, but its detection rule did not find it",
+    "subject": {
+      "kind": "app",
+      "key": "{{app_with_detection_failure.appId}}",
+      "name": "{{app_with_detection_failure.appName}}"
+    },
+    "facts": [
+      {
+        "label": "ESP error",
+        "value": "{{esp_failure_errorcode.value}}",
+        "format": "errorCode"
+      },
+      {
+        "label": "Detection",
+        "value": "{{app_with_detection_failure.value}}"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "detection",

rules/analyze/ANALYZE-APP-014.json

@@ -5,7 +5,7 @@
   "description": "Detects when the Enrollment Status Page failed on the Apps subcategory and the agent's AppX deployment log scan identified a suspected MSIX/Store package. Store/MSIX apps install outside the Win32 pipeline, so the regular per-app tracking sees nothing — this correlation names the likely culprit.",
   "severity": "warning",
   "category": "apps",
-  "version": "1.2.1",
+  "version": "1.3.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -123,6 +123,36 @@
       "url": "https://learn.microsoft.com/en-us/mem/intune/apps/store-apps-microsoft"
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "A Store or MSIX app failed during the ESP: {{appx_candidate.value}}",
+    "subject": {
+      "kind": "app",
+      "key": "{{appx_candidate.value}}",
+      "name": "{{appx_candidate.value}}"
+    },
+    "facts": [
+      {
+        "label": "ESP error",
+        "value": "{{appx_esp_errorcode.value}}",
+        "format": "errorCode"
+      },
+      {
+        "label": "ESP phase",
+        "value": "{{appx_esp_category.value}}"
+      },
+      {
+        "label": "Package",
+        "value": "{{appx_top_candidate_package.value}}",
+        "format": "code"
+      },
+      {
+        "label": "Candidates",
+        "value": "{{appx_candidate_count.value}}",
+        "format": "number"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "appx",

rules/analyze/ANALYZE-APP-015.json

@@ -5,7 +5,7 @@
   "description": "The IME reported its user-session completion line, but at least one required app stayed pending — without ever failing or even starting to install — so the agent kept deferring the enrollment completion signal. Typically an assignment/requirement problem (requirement rules never met, dependency never satisfied, or the app needs a context that is not available during ESP, e.g. a VPN client waiting for user sign-in to a portal).",
   "severity": "high",
   "category": "apps",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "trigger": "correlation",
@@ -76,5 +76,20 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "cause",
+    "summary": "Completion waited on {{apps_pending.value}} pending app(s)",
+    "facts": [
+      {
+        "label": "Pending apps",
+        "value": "{{apps_pending.value}}",
+        "format": "number"
+      },
+      {
+        "label": "Apps",
+        "value": "{{pending_app_names.value}}"
+      }
+    ]
+  },
   "tags": ["apps", "esp", "completion", "assignment", "pending"]
 }

rules/analyze/ANALYZE-APP-016.json

@@ -5,7 +5,7 @@
   "description": "One or more required user-ESP apps never started installing while the ESP AccountSetup apps gate waited on them — the agent observed the ESP page exit with those apps still in a non-terminal, never-started state. Unlike ANALYZE-APP-015 (which reports that completion was DEFERRED because of a pending app, and stays silent once the session reaches a terminal verdict) this rule reports the delivery gap itself and fires regardless of the final outcome: an enrollment that Windows called done while required apps never ran is a successful enrollment with missing software, and the operator must see it.",
   "severity": "warning",
   "category": "apps",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "evaluateOn": [
@@ -102,6 +102,26 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "cause",
+    "summary": "Required apps never started installing; first: {{appName.value}}",
+    "subject": {
+      "kind": "app",
+      "key": "{{appName.appId}}",
+      "name": "{{appName.value}}"
+    },
+    "facts": [
+      {
+        "label": "First app",
+        "value": "{{appName.value}}"
+      },
+      {
+        "label": "State",
+        "value": "{{state.value}}",
+        "format": "code"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "esp",

rules/analyze/ANALYZE-APP-017.json

@@ -5,7 +5,7 @@
   "description": "Detects Win32 app installs that failed because the Windows Installer was busy with another installation (HRESULT 0x80070652 / MSI exit code 1618, ERROR_INSTALL_ALREADY_RUNNING). The IME retries on a fixed budget (~15-17 minutes across several attempts) and then reports the app as failed; when the app is ESP-blocking, the Device Setup Apps subcategory terminates with it.",
   "severity": "high",
   "category": "apps",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "baseConfidence": 70,
@@ -81,6 +81,27 @@
       "url": "https://learn.microsoft.com/en-us/troubleshoot/mem/intune/app-management/troubleshoot-win32-app-install"
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "{{mutex_blocked_install.appName}} was blocked by another installation",
+    "subject": {
+      "kind": "app",
+      "key": "{{mutex_blocked_install.appId}}",
+      "name": "{{mutex_blocked_install.appName}}"
+    },
+    "facts": [
+      {
+        "label": "Exit code",
+        "value": "{{installer_exit_1618.value}}",
+        "format": "code"
+      },
+      {
+        "label": "ESP error",
+        "value": "{{esp_failed_with_mutex_code.value}}",
+        "format": "errorCode"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "msiexec",

rules/analyze/ANALYZE-APP-018.json

@@ -5,7 +5,7 @@
   "description": "Detects Win32 apps whose installer exit code maps to Soft reboot or Hard reboot in the app's Intune return-code table (the '[Win32App] lpExitCode is defined as ...' line). Hard reboot stops app processing and restarts the device in Device Setup - the mid-enrollment restart and second sign-in; Soft reboot leaves the app pending until the next restart while the ESP moves on. Confidence rises for Hard reboot and when a reboot was actually observed in the session.",
   "severity": "warning",
   "category": "apps",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "evaluateOn": [
@@ -86,6 +86,31 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "context",
+    "summary": "{{reboot_return_code.appName}} asked for a restart",
+    "subject": {
+      "kind": "app",
+      "key": "{{reboot_return_code.appId}}",
+      "name": "{{reboot_return_code.appName}}"
+    },
+    "facts": [
+      {
+        "label": "Return code",
+        "value": "{{reboot_return_code.exitCode}}",
+        "format": "code"
+      },
+      {
+        "label": "Return code class",
+        "value": "{{reboot_return_code.value}}"
+      },
+      {
+        "label": "Restarts",
+        "value": "{{reboot_observed.count}}",
+        "format": "number"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "reboot",

rules/analyze/ANALYZE-CORR-003.json

@@ -5,7 +5,7 @@
   "description": "Correlates an active proxy configuration with app content download failures (Delivery Optimization timeout or download error), suggesting the proxy is blocking or throttling Intune content delivery.",
   "severity": "high",
   "category": "apps",
-  "version": "1.2.0",
+  "version": "1.3.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -67,6 +67,30 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "App downloads failed while a {{proxy_active.value}} proxy was configured",
+    "subject": {
+      "kind": "network",
+      "key": "proxy",
+      "name": "Proxy"
+    },
+    "facts": [
+      {
+        "label": "Proxy type",
+        "value": "{{proxy_active.value}}"
+      },
+      {
+        "label": "Failed app",
+        "value": "{{download_failure.appName}}"
+      },
+      {
+        "label": "Error",
+        "value": "{{download_failure.errorCode}}",
+        "format": "errorCode"
+      }
+    ]
+  },
   "tags": [
     "correlation",
     "apps",

rules/analyze/ANALYZE-DEV-001.json

@@ -5,7 +5,7 @@
   "description": "Detects when Windows Hello for Business provisioning did not start within the expected timeframe after ESP completion.",
   "severity": "warning",
   "category": "device",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "enabled": true,
   "baseConfidence": 60,
   "conditions": [
@@ -30,5 +30,21 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "Windows Hello provisioning did not start in time",
+    "subject": {
+      "kind": "device",
+      "key": "windows-hello",
+      "name": "Windows Hello"
+    },
+    "facts": [
+      {
+        "label": "Timed out",
+        "value": "{{hello_timeout.timestamp}}",
+        "format": "timestamp"
+      }
+    ]
+  },
   "tags": ["device", "hello", "timeout"]
 }

rules/analyze/ANALYZE-DEV-002.json

@@ -5,7 +5,7 @@
   "description": "Detects when memory usage stays above 90% across at least 3 performance snapshots during enrollment, indicating sustained memory pressure that can cause app installation failures or system instability.",
   "severity": "warning",
   "category": "device",
-  "version": "2.1.0",
+  "version": "2.2.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -65,6 +65,22 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "context",
+    "summary": "Memory use was above 90% in {{sustained_high_memory.count}} snapshots",
+    "subject": {
+      "kind": "device",
+      "key": "memory",
+      "name": "Memory"
+    },
+    "facts": [
+      {
+        "label": "Snapshots above 90%",
+        "value": "{{sustained_high_memory.count}}",
+        "format": "number"
+      }
+    ]
+  },
   "tags": [
     "device",
     "memory",

rules/analyze/ANALYZE-DEV-003.json

@@ -5,7 +5,7 @@
   "description": "Detects devices enrolling with an OS build that has reached end of support: Windows 11 22H2 (build 22621) or older, which includes every Windows 10 build. Maintenance note: raise the build cutoff to 22631 after Windows 11 23H2 reaches end of support on 2026-11-10.",
   "severity": "warning",
   "category": "device",
-  "version": "2.1.0",
+  "version": "2.2.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -56,6 +56,26 @@
       "url": "https://learn.microsoft.com/en-us/lifecycle/products/windows-11-home-and-pro"
     }
   ],
+  "presentation": {
+    "role": "advisory",
+    "summary": "Windows build {{old_build.value}} is out of support",
+    "subject": {
+      "kind": "device",
+      "key": "os",
+      "name": "Windows version"
+    },
+    "facts": [
+      {
+        "label": "OS",
+        "value": "{{os_version.value}}"
+      },
+      {
+        "label": "Build",
+        "value": "{{old_build.value}}",
+        "format": "code"
+      }
+    ]
+  },
   "tags": [
     "device",
     "os-version",

rules/analyze/ANALYZE-DEV-004.json

@@ -5,7 +5,7 @@
   "description": "Detects a Windows quality/cumulative update that FAILED to install or download during the OOBE / ESP enrollment window (WindowsUpdateClient EventID 20 from the System log, 31 from the Operational channel). A cumulative update failing mid-OOBE is a known cause of enrollment errors and is otherwise invisible in the Intune console.",
   "severity": "high",
   "category": "device",
-  "version": "1.2.0",
+  "version": "1.3.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -113,6 +113,27 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "cause",
+    "summary": "A Windows update failed during setup: {{wu_update_title.value}}",
+    "subject": {
+      "kind": "device",
+      "key": "windows-update",
+      "name": "Windows Update"
+    },
+    "facts": [
+      {
+        "label": "Error",
+        "value": "{{wu_hresult.value}}",
+        "format": "errorCode"
+      },
+      {
+        "label": "Symbol",
+        "value": "{{wu_hresult_symbol.value}}",
+        "format": "code"
+      }
+    ]
+  },
   "tags": [
     "device",
     "windows-update",

rules/analyze/ANALYZE-DEV-005.json

@@ -5,7 +5,7 @@
   "description": "Informational: a Windows quality/cumulative update installed during the OOBE / ESP enrollment window (WindowsUpdateClient EventID 19). Surfaces updates applied mid-enrollment — useful context when correlating enrollment duration, reboots, or later issues with a specific KB.",
   "severity": "info",
   "category": "device",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -61,6 +61,22 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "context",
+    "summary": "A Windows update installed during setup: {{wu_update_title.value}}",
+    "subject": {
+      "kind": "device",
+      "key": "windows-update",
+      "name": "Windows Update"
+    },
+    "facts": [
+      {
+        "label": "Installed",
+        "value": "{{wu_installed.timestamp}}",
+        "format": "timestamp"
+      }
+    ]
+  },
   "tags": [
     "device",
     "windows-update",

rules/analyze/ANALYZE-DEV-006.json

@@ -5,7 +5,7 @@
   "description": "Informational: the OS build (CurrentBuild.UBR) changed across a reboot during enrollment — an update was installed, regardless of which servicing path delivered it. Deterministic corroboration that works even when the Windows Update event channel shows nothing (registry comparison across restart cannot be lost to timing or logging-path differences).",
   "severity": "info",
   "category": "device",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -69,6 +69,27 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "context",
+    "summary": "Windows was updated from build {{previous_build.value}} to {{current_build.value}}",
+    "subject": {
+      "kind": "device",
+      "key": "os",
+      "name": "Windows version"
+    },
+    "facts": [
+      {
+        "label": "Build before",
+        "value": "{{previous_build.value}}",
+        "format": "code"
+      },
+      {
+        "label": "Build after",
+        "value": "{{current_build.value}}",
+        "format": "code"
+      }
+    ]
+  },
   "tags": [
     "device",
     "windows-update",

rules/analyze/ANALYZE-DEV-007.json

@@ -5,7 +5,7 @@
   "description": "The device's real-time clock deviates from NTP by more than two minutes. Microsoft documents that a clock off by several minutes causes TPM attestation errors (self-deploying/pre-provisioning) and ESP timeouts.",
   "severity": "warning",
   "category": "device",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "evaluateOn": [
@@ -17,7 +17,7 @@
   "confidenceThreshold": 50,
   "conditions": [
     {
-      "signal": "clock_ahead",
+      "signal": "clock_behind",
       "source": "event_data",
       "eventType": "ntp_time_check",
       "dataField": "offsetSeconds",
@@ -26,7 +26,7 @@
       "required": false
     },
     {
-      "signal": "clock_behind",
+      "signal": "clock_ahead",
       "source": "event_data",
       "eventType": "ntp_time_check",
       "dataField": "offsetSeconds",
@@ -59,6 +59,26 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "cause",
+    "subject": {
+      "kind": "device",
+      "key": "clock",
+      "name": "System clock"
+    },
+    "facts": [
+      {
+        "label": "Clock offset (s)",
+        "value": "{{clock_behind.value}}",
+        "format": "number"
+      },
+      {
+        "label": "Clock offset (s)",
+        "value": "{{clock_ahead.value}}",
+        "format": "number"
+      }
+    ]
+  },
   "tags": [
     "device",
     "ntp",

rules/analyze/ANALYZE-DEV-008.json

@@ -5,7 +5,7 @@
   "description": "The device's clock changed by five or more minutes mid-session (clock jump) or ran five or more minutes off server time for the whole session (sustained offset), measured against the server receive frame of every uploaded event. Complements ANALYZE-DEV-007, which is a one-shot NTP check at agent startup and misses devices where NTP was unreachable or the clock changed later. IME-log-derived events are excluded from the measurement.",
   "severity": "warning",
   "category": "device",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "evaluateOn": [
@@ -57,6 +57,26 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "cause",
+    "subject": {
+      "kind": "device",
+      "key": "clock",
+      "name": "System clock"
+    },
+    "facts": [
+      {
+        "label": "Clock jump (min)",
+        "value": "{{clock_jump.jumpMinutes}}",
+        "format": "number"
+      },
+      {
+        "label": "Clock offset (min)",
+        "value": "{{sustained_offset.offsetMinutes}}",
+        "format": "number"
+      }
+    ]
+  },
   "tags": [
     "device",
     "clock-skew",

rules/analyze/ANALYZE-DEV-009.json

@@ -5,7 +5,7 @@
   "description": "Detects that the device's battery dropped to 15% or less while enrolling on battery power (power_state_change threshold event from the agent's live power watcher). A device dying mid-enrollment leaves a half-provisioned state that typically needs a full reset; this rule surfaces the risk while there is still time to plug the device in. Uses a single required condition on the agent-stamped thresholdPercent field (the data field 'transition' on the event is unrelated to this rule's 'trigger' property).",
   "severity": "high",
   "category": "device",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -39,6 +39,27 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "context",
+    "summary": "The battery fell below {{battery_threshold_15.value}}% during setup",
+    "subject": {
+      "kind": "device",
+      "key": "power",
+      "name": "Power"
+    },
+    "facts": [
+      {
+        "label": "Battery below (%)",
+        "value": "{{battery_threshold_15.value}}",
+        "format": "number"
+      },
+      {
+        "label": "Reported",
+        "value": "{{battery_threshold_15.timestamp}}",
+        "format": "timestamp"
+      }
+    ]
+  },
   "tags": [
     "device",
     "power",

rules/analyze/ANALYZE-DEV-010.json

@@ -5,7 +5,7 @@
   "description": "Detects that the device lost AC power and switched to battery while the enrollment was running (power_state_change event with transition ac_to_battery from the agent's live power watcher). Uses a single required condition so the match is pinned to one event instance; the event payload carries the battery percentage at the moment of the switch. The agent debounces power flapping (5s) so a brief unplug/replug does not produce this event.",
   "severity": "warning",
   "category": "device",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -39,6 +39,22 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "context",
+    "summary": "The device switched from AC to battery power during setup",
+    "subject": {
+      "kind": "device",
+      "key": "power",
+      "name": "Power"
+    },
+    "facts": [
+      {
+        "label": "Switched",
+        "value": "{{switched_to_battery.timestamp}}",
+        "format": "timestamp"
+      }
+    ]
+  },
   "tags": [
     "device",
     "power",

rules/analyze/ANALYZE-DEV-012.json

@@ -5,7 +5,7 @@
   "description": "Detects that the device completed a sleep, hibernate or Modern Standby episode of at least 5 minutes while the enrollment was running (system_sleep_episode event from the agent's system timeline watcher, recorded retroactively at wake from the Windows event log). The watcher also backfills pre-agent episodes from the last 24 h for timeline context (backfilled=true); those ended before the enrollment started and are excluded via the same-event filter backfilled equals false — a separate condition could not do that, because each condition scans all events on its own and a pre-session hibernate would still satisfy the duration test. A single required condition keeps the match pinned to one event instance so the interpolated duration is the matched value. Short Modern Standby screen-off scenarios (under 60 s of real sleep) never emit the event, and this rule adds a 5-minute floor so brief lid-close moments do not fire it.",
   "severity": "info",
   "category": "device",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -42,6 +42,27 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "context",
+    "summary": "The device went to sleep during setup",
+    "subject": {
+      "kind": "device",
+      "key": "power",
+      "name": "Power"
+    },
+    "facts": [
+      {
+        "label": "Asleep for",
+        "value": "{{slept_during_enrollment.value}}",
+        "format": "duration"
+      },
+      {
+        "label": "Recorded",
+        "value": "{{slept_during_enrollment.timestamp}}",
+        "format": "timestamp"
+      }
+    ]
+  },
   "tags": [
     "device",
     "power",

rules/analyze/ANALYZE-DEV-013.json

@@ -0,0 +1,134 @@
+{
+  "$schema": "../schema/analyze-rule.schema.json",
+  "ruleId": "ANALYZE-DEV-013",
+  "title": "OOBE Quality Update Failed",
+  "description": "Detects that the Windows quality update page at the end of OOBE (Enrollment Status Page setting \"Install Windows quality updates\") reported a failed download or install and recorded no successful install afterwards — the page's own failure names on the oobe_update_page events. Setup usually continues, so the failure is otherwise invisible: the device finishes OOBE without the monthly security update. An OS build change across a restart (os_build_changed) vetoes the finding, and so does the agent's names_capped marker — once the page's event names were capped, a missing success name proves nothing. A downloadInstallFailureHelper entry alone is no failure: updates that installed fine log it too.",
+  "severity": "warning",
+  "category": "device",
+  "version": "1.1.0",
+  "enabled": true,
+  "evaluateOn": [
+    "enrollment_end"
+  ],
+  "trigger": "single",
+  "baseConfidence": 75,
+  "confidenceThreshold": 60,
+  "conditions": [
+    {
+      "signal": "update_failure",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "name",
+      "operator": "in",
+      "value": "SdxWebAppCloudNDUP_processStatusChangeFromHandler_downloadFailedError,SdxWebAppCloudNDUP_processStatusChangeFromHandler_installFailedError,ExpeditedUpdate_commitExpeditionDownloadInstallAsyncFailure",
+      "required": true
+    },
+    {
+      "signal": "no_install_success",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "name",
+      "operator": "not_exists",
+      "value": "",
+      "filterField": "name",
+      "filterOperator": "regex",
+      "filterValue": "installSucceededRebootRequired|commitExpeditionDownloadInstallAsyncSucceeded",
+      "required": true
+    },
+    {
+      "signal": "os_build_unchanged",
+      "source": "event_type",
+      "eventType": "os_build_changed",
+      "operator": "not_exists",
+      "value": "",
+      "required": true
+    },
+    {
+      "signal": "page_names_complete",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "cxhEvent",
+      "operator": "not_exists",
+      "value": "",
+      "filterField": "cxhEvent",
+      "filterOperator": "equals",
+      "filterValue": "names_capped",
+      "required": true
+    },
+    {
+      "signal": "update_page_failed",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "result",
+      "operator": "equals",
+      "value": "fail",
+      "filterField": "page",
+      "filterOperator": "equals",
+      "filterValue": "OobeNDUP",
+      "required": false
+    },
+    {
+      "signal": "update_skipped",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "name",
+      "operator": "regex",
+      "value": "SkipDownloadInstallButtonClicked$",
+      "required": false
+    }
+  ],
+  "confidenceFactors": [
+    {
+      "signal": "update_page_failed",
+      "condition": "exists",
+      "weight": 10
+    }
+  ],
+  "explanation": "The Windows quality update page at the end of OOBE reported `{{update_failure}}`, and no successful install was recorded afterwards. When the download or install of the update fails, the page gives up and setup continues, so the enrollment itself can still succeed — but the device most likely finished OOBE without the monthly security update and installs it only later, through the Windows Update policy that applies after enrollment. Selecting Skip on the page after the failure leads to the same result.",
+  "remediation": [
+    {
+      "title": "Find out why the update failed",
+      "steps": [
+        "Open the oobe_update_page events around the failure in the session timeline: the download progress before it shows how far the download got, and a Skip after it shows that someone chose to continue.",
+        "Check the network path the device uses during OOBE for Windows Update and Delivery Optimization: proxy, TLS inspection and firewall rules for the Windows Update endpoints.",
+        "Compare with other devices: the same failure on many devices of one site points at the network path, on one model at the device, and across the fleet at the update itself."
+      ]
+    },
+    {
+      "title": "Make sure the device gets the update",
+      "steps": [
+        "Verify the device's patch level after enrollment; the update ring assigned to the device installs the missing monthly security update on its normal schedule.",
+        "Where the OOBE quality update keeps failing for one network or device group, decide per Enrollment Status Page profile whether \"Install Windows quality updates\" should stay on."
+      ]
+    }
+  ],
+  "relatedDocs": [
+    {
+      "title": "Microsoft Learn: Install Windows quality updates during OOBE (Enrollment Status Page)",
+      "url": "https://learn.microsoft.com/en-us/intune/device-enrollment/windows/setup-status-page#windows-monthly-security-update-release-details"
+    }
+  ],
+  "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "context",
+    "summary": "The Windows quality update at the end of OOBE failed",
+    "subject": {
+      "kind": "device",
+      "key": "oobe-update",
+      "name": "OOBE quality update"
+    },
+    "facts": [
+      {
+        "label": "Failed step",
+        "value": "{{update_failure.value}}",
+        "format": "code"
+      }
+    ]
+  },
+  "tags": [
+    "device",
+    "windows-update",
+    "oobe",
+    "quality-update"
+  ]
+}

rules/analyze/ANALYZE-DEV-014.json

@@ -0,0 +1,140 @@
+{
+  "$schema": "../schema/analyze-rule.schema.json",
+  "ruleId": "ANALYZE-DEV-014",
+  "title": "OOBE Quality Update Not Offered After an Earlier Failure",
+  "description": "Detects that the Windows quality update page at the end of OOBE found no update to install although the device is out of date, on a device whose update page counts an earlier failed download or install (oobe_update_page: NDUPDownloadInstallPreviousFailureCount above 0, osUpdatesWithNDUPPI false, getDaysOutOfDateSucceeded above 0). No failure and no install success are recorded in this session — a failure in this session is ANALYZE-DEV-013 — and the agent's names_capped marker vetoes the finding, because a missing name proves nothing once the page's names were capped. The device finishes OOBE without the monthly security update.",
+  "severity": "warning",
+  "category": "device",
+  "version": "1.1.0",
+  "enabled": true,
+  "evaluateOn": [
+    "enrollment_end"
+  ],
+  "trigger": "single",
+  "baseConfidence": 75,
+  "confidenceThreshold": 60,
+  "conditions": [
+    {
+      "signal": "previous_failures",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "value",
+      "operator": "gt",
+      "value": "0",
+      "filterField": "name",
+      "filterOperator": "equals",
+      "filterValue": "SdxWebAppCloudNDUP_initialize_NDUPDownloadInstallPreviousFailureCount",
+      "required": true
+    },
+    {
+      "signal": "nothing_offered",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "value",
+      "operator": "equals",
+      "value": "false",
+      "filterField": "name",
+      "filterOperator": "equals",
+      "filterValue": "SdxWebAppCloudNDUP_updateManager_processUpdatesMetadata_osUpdatesWithNDUPPI",
+      "required": true
+    },
+    {
+      "signal": "days_out_of_date",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "value",
+      "operator": "gt",
+      "value": "0",
+      "filterField": "name",
+      "filterOperator": "equals",
+      "filterValue": "ExpeditedUpdate_getDaysOutOfDateSucceeded",
+      "required": true
+    },
+    {
+      "signal": "no_update_failure",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "name",
+      "operator": "not_exists",
+      "value": "",
+      "filterField": "name",
+      "filterOperator": "in",
+      "filterValue": "SdxWebAppCloudNDUP_processStatusChangeFromHandler_downloadFailedError,SdxWebAppCloudNDUP_processStatusChangeFromHandler_installFailedError,ExpeditedUpdate_commitExpeditionDownloadInstallAsyncFailure",
+      "required": true
+    },
+    {
+      "signal": "no_install_success",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "name",
+      "operator": "not_exists",
+      "value": "",
+      "filterField": "name",
+      "filterOperator": "regex",
+      "filterValue": "installSucceededRebootRequired|commitExpeditionDownloadInstallAsyncSucceeded",
+      "required": true
+    },
+    {
+      "signal": "page_names_complete",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "cxhEvent",
+      "operator": "not_exists",
+      "value": "",
+      "filterField": "cxhEvent",
+      "filterOperator": "equals",
+      "filterValue": "names_capped",
+      "required": true
+    }
+  ],
+  "explanation": "The Windows quality update page at the end of OOBE checked for updates and found none to install, although Windows reports the device as {{days_out_of_date}} days out of date. The page's failure counter for this device stands at {{previous_failures}}: an earlier download or install of an OOBE quality update on this device failed. The device finished OOBE without the monthly security update and installs it only later, through the Windows Update policy that applies after enrollment.",
+  "remediation": [
+    {
+      "title": "Check the earlier attempt",
+      "steps": [
+        "Look at the device's earlier enrollment sessions — the failed attempt and its ANALYZE-DEV-013 finding are usually there.",
+        "Verify the device's patch level after enrollment; the update ring assigned to the device installs the missing monthly security update on its normal schedule."
+      ]
+    },
+    {
+      "title": "Reduce how far behind devices start",
+      "steps": [
+        "Devices that start OOBE many months behind need a long update during setup; a current OS image or factory build shortens it and gives a failed attempt less to repeat."
+      ]
+    }
+  ],
+  "relatedDocs": [
+    {
+      "title": "Microsoft Learn: Install Windows quality updates during OOBE (Enrollment Status Page)",
+      "url": "https://learn.microsoft.com/en-us/intune/device-enrollment/windows/setup-status-page#windows-monthly-security-update-release-details"
+    }
+  ],
+  "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "context",
+    "summary": "No quality update was offered although the device is {{days_out_of_date.value}} days out of date",
+    "subject": {
+      "kind": "device",
+      "key": "oobe-update",
+      "name": "OOBE quality update"
+    },
+    "facts": [
+      {
+        "label": "Days out of date",
+        "value": "{{days_out_of_date.value}}",
+        "format": "number"
+      },
+      {
+        "label": "Earlier update failures",
+        "value": "{{previous_failures.value}}",
+        "format": "number"
+      }
+    ]
+  },
+  "tags": [
+    "device",
+    "windows-update",
+    "oobe",
+    "quality-update"
+  ]
+}

rules/analyze/ANALYZE-DEV-015.json

@@ -0,0 +1,133 @@
+{
+  "$schema": "../schema/analyze-rule.schema.json",
+  "ruleId": "ANALYZE-DEV-015",
+  "title": "OOBE Update Check Ended Without a Result",
+  "description": "Detects that the Windows quality update page at the end of OOBE started its Windows Update scan but ended without a scan result (oobe_update_page: startWUScanStartedSuccedded, an OobeNDUP stop and no ExpeditedUpdate_getUpdateResults name), on a device that is out of date and allowed to update. In the observed sessions the page ended two minutes after the scan started, on devices more than a year behind. The device finishes OOBE without the monthly security update. The agent's names_capped marker vetoes the finding: once the page's event names were capped, a missing scan result proves nothing.",
+  "severity": "warning",
+  "category": "device",
+  "version": "1.1.0",
+  "enabled": true,
+  "evaluateOn": [
+    "enrollment_end"
+  ],
+  "trigger": "single",
+  "baseConfidence": 75,
+  "confidenceThreshold": 60,
+  "conditions": [
+    {
+      "signal": "update_allowed",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "value",
+      "operator": "equals",
+      "value": "true",
+      "filterField": "name",
+      "filterOperator": "equals",
+      "filterValue": "ExpeditedUpdate_isNDUPAllowedByCSPSucceeded",
+      "required": true
+    },
+    {
+      "signal": "scan_started",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "name",
+      "operator": "equals",
+      "value": "ExpeditedUpdate_startWUScanStartedSuccedded",
+      "required": true
+    },
+    {
+      "signal": "no_scan_result",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "name",
+      "operator": "not_exists",
+      "value": "",
+      "filterField": "name",
+      "filterOperator": "regex",
+      "filterValue": "^ExpeditedUpdate_getUpdateResults",
+      "required": true
+    },
+    {
+      "signal": "update_page_ended",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "cxhEvent",
+      "operator": "equals",
+      "value": "page_stopped",
+      "filterField": "page",
+      "filterOperator": "equals",
+      "filterValue": "OobeNDUP",
+      "required": true
+    },
+    {
+      "signal": "days_out_of_date",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "value",
+      "operator": "gt",
+      "value": "0",
+      "filterField": "name",
+      "filterOperator": "equals",
+      "filterValue": "ExpeditedUpdate_getDaysOutOfDateSucceeded",
+      "required": true
+    },
+    {
+      "signal": "page_names_complete",
+      "source": "event_data",
+      "eventType": "oobe_update_page",
+      "dataField": "cxhEvent",
+      "operator": "not_exists",
+      "value": "",
+      "filterField": "cxhEvent",
+      "filterOperator": "equals",
+      "filterValue": "names_capped",
+      "required": true
+    }
+  ],
+  "explanation": "The Windows quality update page at the end of OOBE started a Windows Update scan, but no scan result arrived before the page ended. The device is {{days_out_of_date}} days out of date, so an update was due — the device finished OOBE without the monthly security update and installs it only later, through the Windows Update policy that applies after enrollment. The page waits only a short time for the scan (two minutes in the observed sessions): a scan that takes longer, or cannot reach Windows Update, ends the page without an update.",
+  "remediation": [
+    {
+      "title": "Check the path to Windows Update during OOBE",
+      "steps": [
+        "Verify that devices in OOBE reach the Windows Update service: proxy, TLS inspection and firewall rules for the Windows Update endpoints.",
+        "Check whether an update policy applied during the Enrollment Status Page points the scan at another source, such as WSUS."
+      ]
+    },
+    {
+      "title": "Shorten the scan",
+      "steps": [
+        "Devices far behind scan longer; a current OS image or factory build keeps the OOBE scan short.",
+        "Verify the device's patch level after enrollment; the update ring assigned to the device installs the missing monthly security update on its normal schedule."
+      ]
+    }
+  ],
+  "relatedDocs": [
+    {
+      "title": "Microsoft Learn: Install Windows quality updates during OOBE (Enrollment Status Page)",
+      "url": "https://learn.microsoft.com/en-us/intune/device-enrollment/windows/setup-status-page#windows-monthly-security-update-release-details"
+    }
+  ],
+  "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "context",
+    "summary": "The OOBE update check ended without a result",
+    "subject": {
+      "kind": "device",
+      "key": "oobe-update",
+      "name": "OOBE quality update"
+    },
+    "facts": [
+      {
+        "label": "Days out of date",
+        "value": "{{days_out_of_date.value}}",
+        "format": "number"
+      }
+    ]
+  },
+  "tags": [
+    "device",
+    "windows-update",
+    "oobe",
+    "quality-update"
+  ]
+}

rules/analyze/ANALYZE-ENRL-001.json

@@ -5,7 +5,7 @@
   "description": "Detects when the Autopilot enrollment has explicitly failed.",
   "severity": "critical",
   "category": "enrollment",
-  "version": "1.1.2",
+  "version": "1.2.0",
   "enabled": true,
   "baseConfidence": 95,
   "conditions": [
@@ -57,19 +57,40 @@
   "confidenceFactors": [
     { "signal": "failed_app", "condition": "exists", "weight": 5 }
   ],
-  "explanation": "The Autopilot enrollment has explicitly failed.\n\n**Key facts extracted from this session's events:**\n\n- **Failure reason:** `{{reason}}`\n- **First failed app:** `{{appName}}`\n- **ESP failed subcategory:** `{{failedSubcategory}}`\n- **ESP HRESULT:** `{{errorCode}}`\n\nWhen an ESP Apps-subcategory failure is paired with HRESULT `0x87D1041C`, the device-side install most likely completed but the Intune detection rule did not match afterwards — see the **App Detection Failure During ESP** rule (ANALYZE-APP-013) in this same analysis for the actionable root cause and remediation steps. Any field still shown as a literal `{{…}}` placeholder was simply not recorded for this session — for example, a non-ESP failure path has no ESP failed-subcategory or HRESULT.",
+  "explanation": "The Autopilot enrollment has explicitly failed.\n\n**Key facts from this session's events:**\n\n- **Failure reason:** `{{reason}}`\n- **First failed app:** `{{appName}}`\n- **ESP step that failed:** `{{failedSubcategory}}`\n- **ESP error code:** `{{errorCode}}`\n\nA more specific finding in this analysis usually names the cause, such as an app, ESP or timeout finding.",
   "remediation": [
     {
-      "title": "Investigate enrollment failure",
+      "title": "Find the cause",
       "steps": [
-        "Start with the more specific ANALYZE-APP-* / ANALYZE-ESP-* rules that fired alongside this one — they correlate specific HRESULTs to actionable causes (e.g. 0x87D1041C → detection rule mismatch in ANALYZE-APP-013)",
-        "Read the `reason` field on the enrollment_failed event for the immediate trigger (e.g. esp_terminal_failure, hello_timeout, max_lifetime)",
-        "If a HRESULT is present, look it up against the ESP error catalog (Microsoft Learn: Autopilot ESP error codes)",
-        "Review other analyze rule results for correlated issues (disk space, network, app failures)",
-        "Check the device event timeline for the sequence of events leading to failure",
-        "Verify the Autopilot profile and ESP configuration in Intune"
+        "Open the finding in this analysis that names the cause (an app, ESP or timeout finding) and follow its steps",
+        "If no other finding fired, look up the ESP error code and check the failed ESP step in the session timeline",
+        "Verify the Autopilot profile and the ESP settings in Intune"
       ]
     }
   ],
+  "presentation": {
+    "role": "outcome",
+    "summary": "The enrollment failed at the ESP step {{esp_failure_subcategory.value}} with {{esp_failure_errorcode.value}}",
+    "facts": [
+      {
+        "label": "Reason",
+        "value": "{{enrollment_failed_reason.value}}",
+        "format": "code"
+      },
+      {
+        "label": "ESP step",
+        "value": "{{esp_failure_subcategory.value}}"
+      },
+      {
+        "label": "ESP error",
+        "value": "{{esp_failure_errorcode.value}}",
+        "format": "errorCode"
+      },
+      {
+        "label": "First failed app",
+        "value": "{{failed_app.appName}}"
+      }
+    ]
+  },
   "tags": ["enrollment", "failed", "critical"]
 }

rules/analyze/ANALYZE-ENRL-002.json

@@ -2,10 +2,10 @@
   "$schema": "../schema/analyze-rule.schema.json",
   "ruleId": "ANALYZE-ENRL-002",
   "title": "Session Timed Out",
-  "description": "Detects when the enrollment session was force-failed by the server-side inactivity timeout (no agent telemetry for the configured SessionTimeoutHours window) without ever completing or explicitly failing.",
+  "description": "Detects when the server-side inactivity timeout closed the enrollment session (no agent telemetry for the configured SessionTimeoutHours window) before it ever completed or explicitly failed.",
   "severity": "high",
   "category": "enrollment",
-  "version": "1.2.0",
+  "version": "1.4.0",
   "enabled": true,
   "baseConfidence": 80,
   "preconditions": [
@@ -20,6 +20,12 @@
       "eventType": "enrollment_failed",
       "operator": "not_exists",
       "description": "Suppress when an explicit enrollment_failed exists — that path is already covered by ANALYZE-ENRL-001; avoid double-reporting the same session."
+    },
+    {
+      "source": "event_data",
+      "eventType": "device_blocked",
+      "operator": "not_exists",
+      "description": "Suppress when the platform blocked the session's device: the silence is the refused uploads, which ANALYZE-ENRL-006 explains — 'the agent died or the device went offline' would be wrong."
     }
   ],
   "conditions": [
@@ -58,7 +64,7 @@
       "required": false
     }
   ],
-  "explanation": "The enrollment session was force-failed by the server-side maintenance sweep after `{{timeoutHours}}` hour(s) of agent inactivity — the device stopped sending telemetry and never reported completion or an explicit failure.\n\nThis usually means the agent process died or the device went offline mid-enrollment (power off, network loss, bluescreen, reimage) rather than a clean enrollment failure. Because there is no terminal `enrollment_failed` event, the more specific app/ESP rules cannot pinpoint a cause — the evidence simply stops at the last event before the gap.\n\n**If the evidence shows the user desktop was reached and all tracked apps completed**, the device most likely got all its content but the ESP never signaled completion — on hybrid-joined devices the usual culprit is the user session never establishing Entra affinity (Account Setup re-running on every logon). Check whether the hybrid user-affinity finding also fired for this session.",
+  "explanation": "The server-side maintenance sweep closed this session after `{{timeoutHours}}` hour(s) of agent inactivity — the device stopped sending telemetry and never reported completion or an explicit failure.\n\nThis usually means the agent process died or the device went offline mid-enrollment (power off, network loss, bluescreen, reimage) rather than a clean enrollment failure. Because there is no terminal `enrollment_failed` event, the more specific app/ESP rules cannot pinpoint a cause — the evidence simply stops at the last event before the gap.\n\n**If the evidence shows the user desktop was reached and all tracked apps completed**, the device most likely got all its content but the ESP never signaled completion — on hybrid-joined devices the usual culprit is the user session never establishing Entra affinity (Account Setup re-running on every logon). Check whether the hybrid user-affinity finding also fired for this session.",
   "remediation": [
     {
       "title": "Investigate the silent session",
@@ -78,5 +84,21 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "outcome",
+    "summary": "No completion or failure arrived within {{timeoutHours.value}} hours",
+    "facts": [
+      {
+        "label": "Timeout (hours)",
+        "value": "{{timeoutHours.value}}",
+        "format": "number"
+      },
+      {
+        "label": "Desktop reached",
+        "value": "{{desktop_was_reached.timestamp}}",
+        "format": "timestamp"
+      }
+    ]
+  },
   "tags": ["enrollment", "timeout", "stalled"]
 }

rules/analyze/ANALYZE-ENRL-003.json

@@ -5,7 +5,7 @@
   "description": "A TPM-attestation-related error code documented in the Microsoft Autopilot known-issues list was observed (0x800705B4, 0x801C03EA, 0x81039001, 0x81039023, 0x81039024, 0x80070490). These block self-deploying and pre-provisioning deployments during 'Securing your hardware'. Observability: attestation precedes MDM enrollment, so in the normal deployment model (agent installed via IME after enrollment) these codes are rarely visible — the rule primarily serves bootstrap-token deployments.",
   "severity": "high",
   "category": "enrollment",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "evaluateOn": [
@@ -80,6 +80,27 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "cause",
+    "summary": "TPM attestation failed with a known Autopilot error code",
+    "subject": {
+      "kind": "device",
+      "key": "tpm",
+      "name": "TPM attestation"
+    },
+    "facts": [
+      {
+        "label": "Error code",
+        "value": "{{tpm_hresult_esp_failure.value}}",
+        "format": "errorCode"
+      },
+      {
+        "label": "Error code",
+        "value": "{{tpm_hresult_esp_advisory.value}}",
+        "format": "errorCode"
+      }
+    ]
+  },
   "tags": [
     "enrollment",
     "tpm",

rules/analyze/ANALYZE-ENRL-004.json

@@ -5,7 +5,7 @@
   "description": "An enrollment-blocking error code documented in the Microsoft Autopilot known-issues/troubleshooting pages was observed (0x80180014, 0x80180018, 0xC1036501, 0x801C03F3, 0x80070774). Each has a specific, documented cause and fix. Observability: a blocked MDM enrollment normally means no agent gets installed at all (IME install happens after enrollment) — this rule can realistically only fire for bootstrap-token deployments (pre-MDM reporting) or mixed retry scenarios.",
   "severity": "high",
   "category": "enrollment",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "evaluateOn": [
@@ -78,6 +78,27 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "cause",
+    "summary": "MDM enrollment was blocked with a known Autopilot error code",
+    "subject": {
+      "kind": "device",
+      "key": "mdm-enrollment",
+      "name": "MDM enrollment"
+    },
+    "facts": [
+      {
+        "label": "Error code",
+        "value": "{{enrl_hresult_esp_failure.value}}",
+        "format": "errorCode"
+      },
+      {
+        "label": "Error code",
+        "value": "{{enrl_hresult_esp_advisory.value}}",
+        "format": "errorCode"
+      }
+    ]
+  },
   "tags": [
     "enrollment",
     "mdm",

rules/analyze/ANALYZE-ENRL-005.json

@@ -5,7 +5,7 @@
   "description": "A Microsoft Entra hybrid join Autopilot deployment hit error 0x80004005 — a documented build-dependent known issue, resolved by KB5065789 (25H2, build >= 26200.6725), KB5065426 (24H2, build >= 26100.6584), and KB5070312 (23H2, build >= 22631.6276).",
   "severity": "high",
   "category": "enrollment",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "evaluateOn": [
@@ -103,6 +103,27 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "cause",
+    "summary": "Hybrid join hit 0x80004005, a known timeout fixed by Windows updates",
+    "subject": {
+      "kind": "device",
+      "key": "hybrid-join",
+      "name": "Hybrid join"
+    },
+    "facts": [
+      {
+        "label": "Error code",
+        "value": "{{e_fail_esp_failure.value}}",
+        "format": "errorCode"
+      },
+      {
+        "label": "Error code",
+        "value": "{{e_fail_esp_advisory.value}}",
+        "format": "errorCode"
+      }
+    ]
+  },
   "tags": [
     "enrollment",
     "hybrid-join",

rules/analyze/ANALYZE-ENRL-006.json

@@ -0,0 +1,107 @@
+{
+  "$schema": "../schema/analyze-rule.schema.json",
+  "ruleId": "ANALYZE-ENRL-006",
+  "title": "Monitoring Paused by a Protective Limit",
+  "description": "Detects when the platform blocked the session's device — automatically after an unusually high event volume, or by the platform operator — so the session's telemetry stopped reaching Autopilot Monitor.",
+  "severity": "warning",
+  "category": "enrollment",
+  "version": "1.0.0",
+  "author": "Autopilot Monitor",
+  "enabled": true,
+  "evaluateOn": [
+    "enrollment_end",
+    "on_event:device_blocked"
+  ],
+  "trigger": "single",
+  "baseConfidence": 90,
+  "confidenceThreshold": 50,
+  "conditions": [
+    {
+      "signal": "device_blocked",
+      "source": "event_type",
+      "eventType": "device_blocked",
+      "operator": "exists",
+      "value": "",
+      "required": true
+    },
+    {
+      "signal": "event_count",
+      "source": "event_data",
+      "eventType": "device_blocked",
+      "dataField": "eventCount",
+      "operator": "exists",
+      "value": "",
+      "required": false
+    },
+    {
+      "signal": "event_limit",
+      "source": "event_data",
+      "eventType": "device_blocked",
+      "dataField": "threshold",
+      "operator": "exists",
+      "value": "",
+      "required": false
+    },
+    {
+      "signal": "blocked_until",
+      "source": "event_data",
+      "eventType": "device_blocked",
+      "dataField": "blockedUntil",
+      "operator": "exists",
+      "value": "",
+      "required": false
+    }
+  ],
+  "explanation": "Autopilot Monitor stopped accepting telemetry from this session. A protective limit keeps a single device that sends far more events than an enrollment normally does from distorting your fleet data or overloading the platform. When a session crosses it, the device's uploads are paused for a limited time; in severe cases the agent ends monitoring and removes itself. The platform operator can apply the same pause by hand.\n\nA pause does not affect the enrollment itself: the agent keeps running and only holds back its uploads, so Windows Autopilot and Intune carry on as usual. While the pause lasts, no events from this session reach Autopilot Monitor, so the timeline and the final status reflect the last state the agent reported before it. A paused session does not affect the device's next enrollment, which starts unblocked. When the agent ended monitoring, the device stays unmonitored until the end time shown on the session page, new enrollments included.",
+  "remediation": [
+    {
+      "title": "Find what made the session so busy",
+      "steps": [
+        "Open the event timeline and look at which event type repeats most often before the pause — for example a network connection that drops and reconnects every few minutes, or a download that reports progress continuously",
+        "Fix the underlying condition (for example a loose dock, cable or Wi-Fi), then re-enroll the device if you need a complete record of the enrollment",
+        "If no event type stands out, report the session from its page so it can be looked at"
+      ]
+    },
+    {
+      "title": "What happens next",
+      "steps": [
+        "A pause ends on its own at the time shown on the session page; an agent that is still running then resumes uploading",
+        "After a pause, a new enrollment on this device starts unblocked",
+        "After the agent ended monitoring, wait for the end time shown on the session page before re-enrolling if you want the new enrollment monitored"
+      ]
+    }
+  ],
+  "relatedDocs": [
+    {
+      "title": "Autopilot Monitor: Protective Limits",
+      "url": "https://docs.autopilotmonitor.com/concepts/protective-limits"
+    }
+  ],
+  "presentation": {
+    "role": "outcome",
+    "summary": "Monitoring was paused by a protective limit",
+    "facts": [
+      {
+        "label": "Paused at",
+        "value": "{{device_blocked.timestamp}}",
+        "format": "timestamp"
+      },
+      {
+        "label": "Events in this session",
+        "value": "{{event_count.value}}",
+        "format": "number"
+      },
+      {
+        "label": "Event limit",
+        "value": "{{event_limit.value}}",
+        "format": "number"
+      },
+      {
+        "label": "Paused until",
+        "value": "{{blocked_until.value}}",
+        "format": "timestamp"
+      }
+    ]
+  },
+  "tags": ["enrollment", "protective-limit", "blocked"]
+}

rules/analyze/ANALYZE-ESP-001.json

@@ -5,7 +5,7 @@
   "description": "Detects when the DeviceSetup phase of the ESP took longer than 30 minutes, indicating a stuck or very slow blocking app. Confidence increases when the phase exceeds 60 minutes.",
   "severity": "high",
   "category": "esp",
-  "version": "1.2.0",
+  "version": "1.3.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -47,6 +47,27 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "Device setup took {{esp_stalled.durationFormatted}}",
+    "subject": {
+      "kind": "esp",
+      "key": "{{esp_stalled.phase}}",
+      "name": "Device setup"
+    },
+    "facts": [
+      {
+        "label": "Duration",
+        "value": "{{esp_stalled.durationSeconds}}",
+        "format": "duration"
+      },
+      {
+        "label": "Started",
+        "value": "{{esp_stalled.phaseStartTimestamp}}",
+        "format": "timestamp"
+      }
+    ]
+  },
   "tags": [
     "esp",
     "blocking",

rules/analyze/ANALYZE-ESP-002.json

@@ -5,7 +5,7 @@
   "description": "Detects when the Certificates subcategory of the Enrollment Status Page reported a failure, even if the enrollment later appears to have succeeded. Users can click 'Continue anyway' on an ESP error, which masks the underlying certificate provisioning failure — this rule surfaces the transition so it stays visible.",
   "severity": "high",
   "category": "esp",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "evaluateOn": [
@@ -58,6 +58,22 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "cause",
+    "summary": "The ESP certificates step failed",
+    "subject": {
+      "kind": "esp",
+      "key": "Certificates",
+      "name": "Certificates"
+    },
+    "facts": [
+      {
+        "label": "Failed at",
+        "value": "{{certificates_subcategory_failed.timestamp}}",
+        "format": "timestamp"
+      }
+    ]
+  },
   "tags": [
     "esp",
     "certificates",

rules/analyze/ANALYZE-ESP-004.json

@@ -5,7 +5,7 @@
   "description": "Detects an ESP terminal failure where the ESP gave up while a blocking app was still installing (the agent's own esp_apps_timeout verdict — no per-app HRESULT was reported) on a profile that allows 'Continue anyway'. The enrollment is reported as failed, but the user most likely dismissed the error screen and reached the desktop — a soft failure rather than an unusable device. An ESP failure that carries a specific error code (a failed installer, a detection mismatch) is a different case and stays with the rule that explains that code.",
   "severity": "warning",
   "category": "esp",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "enabled": true,
   "baseConfidence": 80,
   "conditions": [
@@ -81,5 +81,30 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "The ESP gave up while {{appName.value}} was still installing ('Continue anyway' allowed)",
+    "subject": {
+      "kind": "app",
+      "key": "{{appName.appId}}",
+      "name": "{{appName.value}}"
+    },
+    "facts": [
+      {
+        "label": "ESP error",
+        "value": "{{enrollment_failed.errorCode}}",
+        "format": "errorCode"
+      },
+      {
+        "label": "ESP time limit (min)",
+        "value": "{{espSyncFailureTimeoutMinutes.value}}",
+        "format": "number"
+      },
+      {
+        "label": "ESP step",
+        "value": "{{failedSubcategory.value}}"
+      }
+    ]
+  },
   "tags": ["esp", "timeout", "continue-anyway", "soft-failure", "enrollment"]
 }

rules/analyze/ANALYZE-ESP-005.json

@@ -5,7 +5,7 @@
   "description": "Attributes an observed enrollment reboot to device-assigned MDM policies that Windows flagged as reboot-required. Fires ONLY when both facts are present in the session: at least one actually-observed reboot (system_reboot_detected) AND reboot-required policy flags (mdm_policy_reboot_required). Policy flags alone — e.g. from a user-policy sync during AccountSetup — force no reboot and stay advisory-silent.",
   "severity": "warning",
   "category": "esp",
-  "version": "1.3.0",
+  "version": "1.4.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "evaluateOn": [
@@ -94,6 +94,31 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "context",
+    "summary": "A device policy forced a restart during setup: {{reboot_forcing_policy_uri.value}}",
+    "subject": {
+      "kind": "policy",
+      "key": "{{reboot_forcing_policy_uri.value}}"
+    },
+    "facts": [
+      {
+        "label": "First policy",
+        "value": "{{reboot_forcing_policy_uri.value}}",
+        "format": "code"
+      },
+      {
+        "label": "Policies",
+        "value": "{{multiple_reboot_uris.value}}",
+        "format": "number"
+      },
+      {
+        "label": "Restarts",
+        "value": "{{reboot_observed.count}}",
+        "format": "number"
+      }
+    ]
+  },
   "tags": [
     "esp",
     "reboot",

rules/analyze/ANALYZE-ESP-006.json

@@ -5,7 +5,7 @@
   "description": "Detects the recovery story: the ESP reported a terminal failure (typically DeviceSetup/Apps), the user most likely pressed 'Try again' on the failure page, the failed step re-ran and completed successfully. The earlier failure did not stick — the enrollment continued past it.",
   "severity": "info",
   "category": "esp",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -98,6 +98,34 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "context",
+    "summary": "The ESP failed in the {{category.value}} phase, then recovered",
+    "subject": {
+      "kind": "esp",
+      "key": "{{failedSubcategory.value}}",
+      "name": "{{failedSubcategory.value}}"
+    },
+    "facts": [
+      {
+        "label": "ESP phase",
+        "value": "{{category.value}}"
+      },
+      {
+        "label": "ESP step",
+        "value": "{{failedSubcategory.value}}"
+      },
+      {
+        "label": "Likely apps",
+        "value": "{{likelyCulpritApps.value}}"
+      },
+      {
+        "label": "Recovered after (min)",
+        "value": "{{minutesSinceFailure.value}}",
+        "format": "number"
+      }
+    ]
+  },
   "tags": [
     "esp",
     "recovery",

rules/analyze/ANALYZE-ID-001.json

@@ -12,7 +12,7 @@
     "enrollment_end",
     "whiteglove_sealed"
   ],
-  "version": "1.3.0",
+  "version": "1.4.0",
   "baseConfidence": 75,
   "conditions": [
     {
@@ -55,6 +55,15 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "advisory",
+    "summary": "The expected machine certificate was not found",
+    "subject": {
+      "kind": "device",
+      "key": "machine-certificate",
+      "name": "Machine certificate"
+    }
+  },
   "tags": [
     "identity",
     "certificates",

rules/analyze/ANALYZE-ID-002.json

@@ -5,7 +5,7 @@
   "description": "Detects when the LocalAdminAnalyzer flags unexpected local user accounts or profile directories on the device. This is a known Autopilot bypass technique where admin accounts are created before or during enrollment to gain persistent local admin access.",
   "severity": "high",
   "category": "identity",
-  "version": "1.3.0",
+  "version": "1.4.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "evaluateOn": [
@@ -41,6 +41,30 @@
       "operator": "equals",
       "value": "error",
       "required": false
+    },
+    {
+      "signal": "unexpected_accounts",
+      "source": "event_data",
+      "eventType": "local_admin_analysis",
+      "dataField": "checks.unexpected_accounts",
+      "operator": "not_equals",
+      "value": "",
+      "filterField": "finding",
+      "filterOperator": "equals",
+      "filterValue": "unexpected_local_admins_detected",
+      "required": false
+    },
+    {
+      "signal": "unexpected_admin_members",
+      "source": "event_data",
+      "eventType": "local_admin_analysis",
+      "dataField": "checks.unexpected_admin_members",
+      "operator": "not_equals",
+      "value": "",
+      "filterField": "finding",
+      "filterOperator": "equals",
+      "filterValue": "unexpected_local_admins_detected",
+      "required": false
     }
   ],
   "confidenceFactors": [
@@ -85,6 +109,26 @@
     }
   ],
   "relatedDocs": [],
+  "presentation": {
+    "role": "advisory",
+    "summary": "Unexpected local accounts: {{unexpected_accounts.value}}",
+    "subject": {
+      "kind": "account",
+      "key": "{{unexpected_accounts.value}}"
+    },
+    "facts": [
+      {
+        "label": "Unexpected accounts",
+        "value": "{{unexpected_accounts.value}}",
+        "format": "code"
+      },
+      {
+        "label": "In Administrators",
+        "value": "{{unexpected_admin_members.value}}",
+        "format": "code"
+      }
+    ]
+  },
   "tags": [
     "identity",
     "security",

rules/analyze/ANALYZE-ID-003.json

@@ -5,7 +5,7 @@
   "description": "Fires when the vulnerability correlation finds critical-severity CVEs or CISA KEV entries in installed software. This indicates the device has software with known, actively exploited vulnerabilities.",
   "severity": "critical",
   "category": "security",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "evaluateOn": [
@@ -72,6 +72,27 @@
       "url": "https://nvd.nist.gov/"
     }
   ],
+  "presentation": {
+    "role": "advisory",
+    "summary": "Installed software has {{critical_vulnerabilities.value}} critical vulnerabilities",
+    "subject": {
+      "kind": "device",
+      "key": "vulnerabilities",
+      "name": "Installed software"
+    },
+    "facts": [
+      {
+        "label": "Critical CVEs",
+        "value": "{{critical_vulnerabilities.value}}",
+        "format": "number"
+      },
+      {
+        "label": "Known exploited",
+        "value": "{{kev_matches.value}}",
+        "format": "number"
+      }
+    ]
+  },
   "tags": [
     "vulnerability",
     "cve",

rules/analyze/ANALYZE-ID-004.json

@@ -5,7 +5,7 @@
   "description": "On a hybrid-joined device the real user reached the desktop, but the Intune Management Extension never acquired a Microsoft Entra user token for that user within the wait window — repeatedly across reboots when the warning recurs. Without a user token the signed-in user has no Entra user affinity: user-targeted apps and policies cannot apply, Windows Hello for Business cannot provision, and with a user ESP the Account Setup page re-runs on every logon.",
   "severity": "high",
   "category": "identity",
-  "version": "2.0.0",
+  "version": "2.1.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "trigger": "correlation",
@@ -44,6 +44,24 @@
       "operator": "not_exists",
       "value": "",
       "required": true
+    },
+    {
+      "signal": "token_failure_count",
+      "source": "event_data",
+      "eventType": "entra_user_affinity_pending",
+      "dataField": "tokenFailureCount",
+      "operator": "exists",
+      "value": "",
+      "required": false
+    },
+    {
+      "signal": "token_failure_codes",
+      "source": "event_data",
+      "eventType": "entra_user_affinity_pending",
+      "dataField": "tokenFailureCodes",
+      "operator": "not_equals",
+      "value": "",
+      "required": false
     }
   ],
   "confidenceFactors": [
@@ -96,5 +114,36 @@
     }
   ],
   "markSessionAsFailedDefault": false,
+  "presentation": {
+    "role": "cause",
+    "summary": "The user reached the desktop, but Entra user affinity never formed",
+    "subject": {
+      "kind": "account",
+      "key": "entra-user-affinity",
+      "name": "Entra user affinity"
+    },
+    "facts": [
+      {
+        "label": "Token failures",
+        "value": "{{token_failure_count.value}}",
+        "format": "number"
+      },
+      {
+        "label": "Token error codes",
+        "value": "{{token_failure_codes.value}}",
+        "format": "code"
+      },
+      {
+        "label": "Reported in agent runs",
+        "value": "{{user_affinity_pending.count}}",
+        "format": "number"
+      },
+      {
+        "label": "Desktop reached",
+        "value": "{{user_desktop_reached.timestamp}}",
+        "format": "timestamp"
+      }
+    ]
+  },
   "tags": ["identity", "hybrid-join", "entra-id", "user-affinity", "esp"]
 }

rules/analyze/ANALYZE-NET-001.json

@@ -5,7 +5,7 @@
   "description": "Detects diagnostics package uploads that were answered by a security appliance's HTML block page instead of the Autopilot Monitor backend. A forward proxy / secure web gateway (SWG) with SSL inspection is intercepting the upload endpoint, so exactly the sessions where diagnostics matter most (OnFailure uploads) are the ones losing them.",
   "severity": "warning",
   "category": "network",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "baseConfidence": 75,
@@ -71,6 +71,15 @@
       "url": "https://learn.microsoft.com/en-us/autopilot/networking-requirements"
     }
   ],
+  "presentation": {
+    "role": "advisory",
+    "summary": "A web security gateway blocked the diagnostics upload",
+    "subject": {
+      "kind": "network",
+      "key": "security-gateway",
+      "name": "Web security gateway"
+    }
+  },
   "tags": [
     "network",
     "proxy",

rules/analyze/ANALYZE-OFFICE-001.json

@@ -5,7 +5,7 @@
   "description": "Detects when the Microsoft 365 Apps (Office Click-to-Run) install failed. The V2 OfficeInstallDetector watches the C2R client directly (registry, Delivery Optimization, process) and emits office_install_failed when the C2R INSTALL scenario records an error code. Completion is proven by the core Office binaries on disk; an install window that closes without that proof ends silently, never as a failure. This catches Office failures the deployment status hides: the Intune built-in app type is marked done once the configuration is delivered, and a Win32 or RealmJoin package once its installer exits, while C2R keeps installing in the background.",
   "severity": "warning",
   "category": "apps",
-  "version": "1.1.1",
+  "version": "1.2.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -52,6 +52,22 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "cause",
+    "summary": "Microsoft 365 Apps failed to install",
+    "subject": {
+      "kind": "app",
+      "key": "office-c2r",
+      "name": "Microsoft 365 Apps"
+    },
+    "facts": [
+      {
+        "label": "Error",
+        "value": "{{error_code_present.value}}",
+        "format": "errorCode"
+      }
+    ]
+  },
   "tags": [
     "apps",
     "office",

rules/analyze/ANALYZE-SEC-001.json

@@ -5,7 +5,7 @@
   "description": "Detects devices where the Windows UEFI CA 2023 certificate has not been deployed to the Secure Boot DB. The 2011 certificates expired in June 2026 — affected devices no longer receive Secure Boot security updates until the 2023 certificate is deployed. Since v3 the agent reads the firmware db/KEK variables directly: when the firmware itself confirms the 2023 certificate, the rule is suppressed even if the Windows servicing status has not caught up yet.",
   "severity": "warning",
   "category": "security",
-  "version": "3.1.0",
+  "version": "3.2.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -132,6 +132,35 @@
       "url": "https://support.microsoft.com/en-us/topic/kb5036210"
     }
   ],
+  "presentation": {
+    "role": "advisory",
+    "summary": "The Windows UEFI CA 2023 certificate is not deployed",
+    "subject": {
+      "kind": "device",
+      "key": "secure-boot",
+      "name": "Secure Boot"
+    },
+    "facts": [
+      {
+        "label": "CA 2023 status",
+        "value": "{{ca2023_not_updated.value}}",
+        "format": "code"
+      },
+      {
+        "label": "Error",
+        "value": "{{ca2023_error.value}}",
+        "format": "code"
+      },
+      {
+        "label": "CA 2023 in firmware DB",
+        "value": "{{db_ca2023_firmware_missing.value}}"
+      },
+      {
+        "label": "KEK 2023 in firmware",
+        "value": "{{kek_2023_missing.value}}"
+      }
+    ]
+  },
   "tags": [
     "security",
     "secureboot",

rules/analyze/ANALYZE-SEC-003.json

@@ -5,7 +5,7 @@
   "description": "Fires when the AutoLogon analyzer reports a DefaultPassword value present in Winlogon. A plaintext AutoLogon password stored in the registry is readable by any local administrator and is a credential exposure — more severe than AutoLogon alone (the Sysinternals Autologon tool deliberately avoids this by storing the secret in the LSA instead).",
   "severity": "high",
   "category": "security",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -49,6 +49,15 @@
       "url": "https://learn.microsoft.com/en-us/sysinternals/downloads/autologon"
     }
   ],
+  "presentation": {
+    "role": "advisory",
+    "summary": "A plaintext AutoLogon password is stored in the registry",
+    "subject": {
+      "kind": "device",
+      "key": "autologon",
+      "name": "AutoLogon"
+    }
+  },
   "tags": [
     "security",
     "autologon",

rules/analyze/ANALYZE-SEC-004.json

@@ -5,7 +5,7 @@
   "description": "Optional kiosk allow-list for AutoLogon. Fires when AutoLogon is enabled for a default user that is NOT on your configured list of approved kiosk / shared-device accounts. Disabled until configured because a normal Autopilot enrollment leaves Windows' own ESP auto-logon for the enrolling user (AutoLogon enabled, a default user, no plaintext password) — so AutoLogon-enabled alone is not graded by default; only a plaintext DefaultPassword on disk (ANALYZE-SEC-003) is. Enable this rule with your expected kiosk account name(s) if you want to surface AutoLogon configured for any other user.",
   "severity": "warning",
   "category": "security",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "enabled": false,
   "evaluateOn": [
     "enrollment_end",
@@ -59,6 +59,21 @@
       "url": "https://learn.microsoft.com/en-us/troubleshoot/windows-server/user-profiles-and-logon/turn-on-automatic-logon"
     }
   ],
+  "presentation": {
+    "role": "advisory",
+    "summary": "AutoLogon is enabled for an unexpected user: {{unexpected_autologon_user.value}}",
+    "subject": {
+      "kind": "account",
+      "key": "{{unexpected_autologon_user.value}}"
+    },
+    "facts": [
+      {
+        "label": "AutoLogon user",
+        "value": "{{unexpected_autologon_user.value}}",
+        "format": "code"
+      }
+    ]
+  },
   "tags": [
     "security",
     "autologon",

rules/analyze/ANALYZE-SEC-005.json

@@ -5,7 +5,7 @@
   "description": "Detects provisioning packages (PPKG) that are not on the built-in allow-list. The agent emits one or more provisioning_package_scan events whose artifacts[] arrays list every PPKG (registry package, .ppkg file, or Recovery\\Customizations residue; a large set is split across several events). This rule iterates that array across all of them and fires for any artifact whose identity does not match the allow-list regex. The default allow-list covers the Windows OS-inbox packages (Cosa, Power.EnergyEstimationEngine, Power.Settings, SecureStart) plus common OEM factory-preload families (Intel PPM power profiles, AMD Power profiles, Lenovo ThinkPad power/battery packages, Microsoft Surface defaults and carrier packages) so clean factory devices stay quiet. OEM imaging leftovers (e.g. USMT.PPKG, InstallOrder.ini) and generically named packages deliberately still alert — on a pure-Autopilot device they are worth a look. To allow your own vendor/recovery packages, enable the template rule ANALYZE-SEC-006, extend its allow-list, then disable this rule.",
   "severity": "warning",
   "category": "security",
-  "version": "2.2.0",
+  "version": "2.3.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -38,6 +38,27 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "advisory",
+    "summary": "Provisioning package not on the allow list: {{ppkg_not_allowlisted.value}}",
+    "subject": {
+      "kind": "device",
+      "key": "provisioning-packages",
+      "name": "Provisioning packages"
+    },
+    "facts": [
+      {
+        "label": "Package",
+        "value": "{{ppkg_not_allowlisted.value}}",
+        "format": "code"
+      },
+      {
+        "label": "Unexpected packages",
+        "value": "{{ppkg_not_allowlisted.matchCount}}",
+        "format": "number"
+      }
+    ]
+  },
   "tags": [
     "security",
     "provisioning-package",

rules/analyze/ANALYZE-SEC-006.json

@@ -5,7 +5,7 @@
   "description": "Template version of ANALYZE-SEC-005 with a customizable allow-list. Iterates the provisioning_package_scan artifacts[] array and fires for any PPKG whose identity is not on YOUR allow-list regex. The placeholder ships with the same OS-inbox defaults as ANALYZE-SEC-005 — extend it with your own vendor/recovery packages. This rule is OFF by default: enable it, set your allow-list, then disable ANALYZE-SEC-005 so only packages outside your list alert.",
   "severity": "warning",
   "category": "security",
-  "version": "2.2.0",
+  "version": "2.3.0",
   "enabled": false,
   "evaluateOn": [
     "enrollment_end",
@@ -45,6 +45,27 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "advisory",
+    "summary": "Provisioning package not on the allow list: {{ppkg_not_allowlisted.value}}",
+    "subject": {
+      "kind": "device",
+      "key": "provisioning-packages",
+      "name": "Provisioning packages"
+    },
+    "facts": [
+      {
+        "label": "Package",
+        "value": "{{ppkg_not_allowlisted.value}}",
+        "format": "code"
+      },
+      {
+        "label": "Unexpected packages",
+        "value": "{{ppkg_not_allowlisted.matchCount}}",
+        "format": "number"
+      }
+    ]
+  },
   "tags": [
     "security",
     "provisioning-package",

rules/analyze/ANALYZE-SEC-007.json

@@ -5,7 +5,7 @@
   "description": "Fires when the provisioning-package scan hit a per-source collection cap (too many .ppkg files, registry packages, or Recovery\\Customizations residue files to capture them all). When that happens the allow-list rules (ANALYZE-SEC-005/006) could not evaluate every package, so a non-allow-listed package may have gone unseen. A device with that many provisioning packages is itself unusual and worth a manual look.",
   "severity": "warning",
   "category": "security",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "enabled": true,
   "evaluateOn": [
     "enrollment_end",
@@ -36,6 +36,15 @@
       ]
     }
   ],
+  "presentation": {
+    "role": "advisory",
+    "summary": "The provisioning package scan hit its capture limit",
+    "subject": {
+      "kind": "device",
+      "key": "provisioning-packages",
+      "name": "Provisioning packages"
+    }
+  },
   "tags": [
     "security",
     "provisioning-package",

rules/analyze/ANALYZE-SEC-008.json

@@ -5,7 +5,7 @@
   "description": "Detects devices enrolling with Secure Boot turned off. Secure Boot is a firmware-level protection against boot-time malware and is a prerequisite for several Windows security features (VBS/Credential Guard, device health attestation) and common Intune compliance policies. Devices that support Secure Boot but have it disabled should be flagged during enrollment. Legacy BIOS devices (no Secure Boot capability) are excluded — the agent marks those with a note field when the SecureBoot registry state key is absent.",
   "severity": "warning",
   "category": "security",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "author": "Autopilot Monitor",
   "enabled": true,
   "evaluateOn": [
@@ -82,6 +82,15 @@
       "url": "https://learn.microsoft.com/en-us/windows-hardware/manufacture/desktop/disabling-secure-boot"
     }
   ],
+  "presentation": {
+    "role": "advisory",
+    "summary": "Secure Boot is disabled",
+    "subject": {
+      "kind": "device",
+      "key": "secure-boot",
+      "name": "Secure Boot"
+    }
+  },
   "tags": [
     "security",
     "secureboot",

rules/schema/analyze-rule.schema.json

@@ -133,6 +133,7 @@
           "items": { "$ref": "#/$defs/relatedDoc" },
           "default": []
         },
+        "presentation": { "$ref": "#/$defs/rulePresentation" },
         "tags": {
           "type": "array",
           "description": "Tags for filtering and categorization.",
@@ -352,6 +353,83 @@
       "required": ["title", "url"],
       "additionalProperties": false
     },
+    "rulePresentation": {
+      "type": "object",
+      "description": "How a finding of this rule is shown: its role, a one-line summary, what it is about and its key facts. Every string is a {{token}} template resolved from the finding's evidence when it is rendered; a fact whose value was not recorded is left out.",
+      "properties": {
+        "role": {
+          "type": "string",
+          "enum": ["outcome", "cause", "context", "advisory"],
+          "description": "outcome: restates the session outcome (shown in the verdict, not as a finding row). cause: can explain why a session failed or is stuck. context: a circumstance such as standby, a restart or battery power. advisory: hygiene or security without bearing on the outcome."
+        },
+        "summary": {
+          "type": "string",
+          "minLength": 1,
+          "maxLength": 160,
+          "description": "One line for the finding row, for example \"{{install_retry_loop.appName}} was started {{install_retry_loop.count}} times\". When a token was not recorded, the first sentence of the explanation is shown instead."
+        },
+        "subject": { "$ref": "#/$defs/ruleSubject" },
+        "facts": {
+          "type": "array",
+          "description": "Key facts shown as label/value pairs, in this order.",
+          "maxItems": 6,
+          "items": { "$ref": "#/$defs/ruleFact" }
+        }
+      },
+      "required": ["role"],
+      "additionalProperties": false
+    },
+    "ruleSubject": {
+      "type": "object",
+      "description": "What the finding is about. Findings with the same kind and resolved key are grouped, for example every finding about one app. Use qualified tokens of one signal for key and name, such as {{failed_app.appId}} and {{failed_app.appName}}, so both come from the same event.",
+      "properties": {
+        "kind": {
+          "type": "string",
+          "enum": ["app", "policy", "account", "script", "device", "network", "esp", "session"],
+          "description": "The kind of thing the key identifies."
+        },
+        "key": {
+          "type": "string",
+          "minLength": 1,
+          "maxLength": 200,
+          "description": "Template of the grouping key."
+        },
+        "name": {
+          "type": "string",
+          "minLength": 1,
+          "maxLength": 200,
+          "description": "Template of the display name; the key is shown when it is missing."
+        }
+      },
+      "required": ["kind", "key"],
+      "additionalProperties": false
+    },
+    "ruleFact": {
+      "type": "object",
+      "description": "One key fact of a finding.",
+      "properties": {
+        "label": {
+          "type": "string",
+          "minLength": 1,
+          "maxLength": 40,
+          "description": "Short label, for example \"Install starts\"."
+        },
+        "value": {
+          "type": "string",
+          "minLength": 1,
+          "maxLength": 200,
+          "description": "Template of the value, for example {{install_retry_loop.count}}."
+        },
+        "format": {
+          "type": "string",
+          "enum": ["text", "code", "errorCode", "duration", "number", "timestamp"],
+          "description": "How the value is shown. errorCode adds the error-code catalog meaning, duration reads seconds, timestamp an ISO 8601 time.",
+          "default": "text"
+        }
+      },
+      "required": ["label", "value"],
+      "additionalProperties": false
+    },
     "templateVariable": {
       "type": "object",
       "description": "A variable in a rule condition that must be customized per-tenant before the rule can be used.",

rules/scripts/.npmrc

@@ -1,3 +0,0 @@
-# ajv-cli declares ts-node as a peer dependency it never needs for JSON schemas; do not
-# pull ts-node + typescript into a job that pushes to main.
-legacy-peer-deps=true

rules/scripts/combine.js

@@ -86,12 +86,12 @@ for (const config of configs) {
 // ── Guardrails generation ───────────────────────────────────────────────────
 
 const guardrailsPath = path.join(rulesRoot, 'guardrails.json');
-// NOTE: consumers import from "@/utils/guardrails.generated" (see guardValidation.ts
+// NOTE: consumers import from "@/lib/generated/guardrails.generated" (see guardValidation.ts
 // and SectionGatherRules.tsx). The web re-org (commit a3e8e044) moved this file to
 // utils/ but left this output path pointing at the now-orphaned lib/ copy, which
 // silently froze the live allowlist. Keep this in sync with the import path.
 const guardrailsOutput = path.resolve(
-  rulesRoot, '..', 'src', 'Web', 'autopilot-monitor-web', 'utils', 'guardrails.generated.ts'
+  rulesRoot, '..', 'src', 'Web', 'autopilot-monitor-web', 'lib', 'generated', 'guardrails.generated.ts'
 );
 
 if (fs.existsSync(guardrailsPath)) {

rules/scripts/package-lock.json

@@ -6,7 +6,7 @@
     "": {
       "name": "autopilot-monitor-rules-tooling",
       "devDependencies": {
-        "ajv-cli": "5.0.0",
+        "ajv": "8.20.0",
         "ajv-formats": "3.0.1"
       }
     },
@@ -27,33 +27,6 @@
         "url": "https://github.com/sponsors/epoberezkin"
       }
     },
-    "node_modules/ajv-cli": {
-      "version": "5.0.0",
-      "resolved": "https://registry.npmjs.org/ajv-cli/-/ajv-cli-5.0.0.tgz",
-      "integrity": "sha512-LY4m6dUv44HTyhV+u2z5uX4EhPYTM38Iv1jdgDJJJCyOOuqB8KtZEGjPZ2T+sh5ZIJrXUfgErYx/j3gLd3+PlQ==",
-      "dev": true,
-      "license": "MIT",
-      "dependencies": {
-        "ajv": "^8.0.0",
-        "fast-json-patch": "^2.0.0",
-        "glob": "^7.1.0",
-        "js-yaml": "^3.14.0",
-        "json-schema-migrate": "^2.0.0",
-        "json5": "^2.1.3",
-        "minimist": "^1.2.0"
-      },
-      "bin": {
-        "ajv": "dist/index.js"
-      },
-      "peerDependencies": {
-        "ts-node": ">=9.0.0"
-      },
-      "peerDependenciesMeta": {
-        "ts-node": {
-          "optional": true
-        }
-      }
-    },
     "node_modules/ajv-formats": {
       "version": "3.0.1",
       "resolved": "https://registry.npmjs.org/ajv-formats/-/ajv-formats-3.0.1.tgz",
@@ -72,55 +45,6 @@
         }
       }
     },
-    "node_modules/argparse": {
-      "version": "1.0.10",
-      "resolved": "https://registry.npmjs.org/argparse/-/argparse-1.0.10.tgz",
-      "integrity": "sha512-o5Roy6tNG4SL/FOkCAN6RzjiakZS25RLYFrcMttJqbdd8BWrnA+fGz57iN5Pb06pvBGvl5gQ0B48dJlslXvoTg==",
-      "dev": true,
-      "license": "MIT",
-      "dependencies": {
-        "sprintf-js": "~1.0.2"
-      }
-    },
-    "node_modules/balanced-match": {
-      "version": "1.0.2",
-      "resolved": "https://registry.npmjs.org/balanced-match/-/balanced-match-1.0.2.tgz",
-      "integrity": "sha512-3oSeUO0TMV67hN1AmbXsK4yaqU7tjiHlbxRDZOpH0KW9+CeX4bRAaX0Anxt0tx2MrpRpWwQaPwIlISEJhYU5Pw==",
-      "dev": true,
-      "license": "MIT"
-    },
-    "node_modules/brace-expansion": {
-      "version": "1.1.21",
-      "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-1.1.21.tgz",
-      "integrity": "sha512-9zeA+KLZNNzglF2TPKRQEDyx6Yby7daAkuy8MiPzpXPsYDWi/DRM8jmwUDxokQjYqBpv5DgPiwD4h4ZZSy1Ujw==",
-      "dev": true,
-      "license": "MIT",
-      "dependencies": {
-        "balanced-match": "^1.0.0",
-        "concat-map": "0.0.1"
-      }
-    },
-    "node_modules/concat-map": {
-      "version": "0.0.1",
-      "resolved": "https://registry.npmjs.org/concat-map/-/concat-map-0.0.1.tgz",
-      "integrity": "sha512-/Srv4dswyQNBfohGpz9o6Yb3Gz3SrUDqBH5rTuhGR7ahtlbYKnVxw2bCFMRljaA7EXHaXZ8wsHdodFvbkhKmqg==",
-      "dev": true,
-      "license": "MIT"
-    },
-    "node_modules/esprima": {
-      "version": "4.0.1",
-      "resolved": "https://registry.npmjs.org/esprima/-/esprima-4.0.1.tgz",
-      "integrity": "sha512-eGuFFw7Upda+g4p+QHvnW0RyTX/SVeJBDM/gCtMARO0cLuT2HcEKnTPvhjV6aGeqrCB/sbNop0Kszm0jsaWU4A==",
-      "dev": true,
-      "license": "BSD-2-Clause",
-      "bin": {
-        "esparse": "bin/esparse.js",
-        "esvalidate": "bin/esvalidate.js"
-      },
-      "engines": {
-        "node": ">=4"
-      }
-    },
     "node_modules/fast-deep-equal": {
       "version": "3.1.3",
       "resolved": "https://registry.npmjs.org/fast-deep-equal/-/fast-deep-equal-3.1.3.tgz",
@@ -128,13 +52,6 @@
       "dev": true,
       "license": "MIT"
     },
-    "node_modules/fast-json-patch": {
-      "version": "3.1.1",
-      "resolved": "https://registry.npmjs.org/fast-json-patch/-/fast-json-patch-3.1.1.tgz",
-      "integrity": "sha512-vf6IHUX2SBcA+5/+4883dsIjpBTqmfBjmYiWK1savxQmFk4JfBMLa7ynTYOs1Rolp/T1betJxHiGD3g1Mn8lUQ==",
-      "dev": true,
-      "license": "MIT"
-    },
     "node_modules/fast-uri": {
       "version": "3.1.8",
       "resolved": "https://registry.npmjs.org/fast-uri/-/fast-uri-3.1.8.tgz",
@@ -152,78 +69,6 @@
       ],
       "license": "BSD-3-Clause"
     },
-    "node_modules/fs.realpath": {
-      "version": "1.0.0",
-      "resolved": "https://registry.npmjs.org/fs.realpath/-/fs.realpath-1.0.0.tgz",
-      "integrity": "sha512-OO0pH2lK6a0hZnAdau5ItzHPI6pUlvI7jMVnxUQRtw4owF2wk8lOSabtGDCTP4Ggrg2MbGnWO9X8K1t4+fGMDw==",
-      "dev": true,
-      "license": "ISC"
-    },
-    "node_modules/glob": {
-      "version": "7.2.3",
-      "resolved": "https://registry.npmjs.org/glob/-/glob-7.2.3.tgz",
-      "integrity": "sha512-nFR0zLpU2YCaRxwoCJvL6UvCH2JFyFVIvwTLsIf21AuHlMskA1hhTdk+LlYJtOlYt9v6dvszD2BGRqBL+iQK9Q==",
-      "deprecated": "Old versions of glob are not supported, and contain widely publicized security vulnerabilities, which have been fixed in the current version. Please update. Support for old versions may be purchased (at exorbitant rates) by contacting i@izs.me",
-      "dev": true,
-      "license": "ISC",
-      "dependencies": {
-        "fs.realpath": "^1.0.0",
-        "inflight": "^1.0.4",
-        "inherits": "2",
-        "minimatch": "^3.1.1",
-        "once": "^1.3.0",
-        "path-is-absolute": "^1.0.0"
-      },
-      "engines": {
-        "node": "*"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/isaacs"
-      }
-    },
-    "node_modules/inflight": {
-      "version": "1.0.6",
-      "resolved": "https://registry.npmjs.org/inflight/-/inflight-1.0.6.tgz",
-      "integrity": "sha512-k92I/b08q4wvFscXCLvqfsHCrjrF7yiXsQuIVvVE7N82W3+aqpzuUdBbfhWcy/FZR3/4IgflMgKLOsvPDrGCJA==",
-      "deprecated": "This module is not supported, and leaks memory. Do not use it. Check out lru-cache if you want a good and tested way to coalesce async requests by a key value, which is much more comprehensive and powerful.",
-      "dev": true,
-      "license": "ISC",
-      "dependencies": {
-        "once": "^1.3.0",
-        "wrappy": "1"
-      }
-    },
-    "node_modules/inherits": {
-      "version": "2.0.4",
-      "resolved": "https://registry.npmjs.org/inherits/-/inherits-2.0.4.tgz",
-      "integrity": "sha512-k/vGaX4/Yla3WzyMCvTQOXYeIHvqOKtnqBduzTHpzpQZzAskKMhZ2K+EnBiSM9zGSoIFeMpXKxa4dYeZIQqewQ==",
-      "dev": true,
-      "license": "ISC"
-    },
-    "node_modules/js-yaml": {
-      "version": "3.15.2",
-      "resolved": "https://registry.npmjs.org/js-yaml/-/js-yaml-3.15.2.tgz",
-      "integrity": "sha512-6EuL879VkRA+1Cz578mKMiKvjPNEuk6+r1JaFzoSWejZmtf7xWbIyw1e3KkxlkzTIt9Taw6JBhEppG7utc1P+w==",
-      "dev": true,
-      "license": "MIT",
-      "dependencies": {
-        "argparse": "^1.0.7",
-        "esprima": "^4.0.0"
-      },
-      "bin": {
-        "js-yaml": "bin/js-yaml.js"
-      }
-    },
-    "node_modules/json-schema-migrate": {
-      "version": "2.0.0",
-      "resolved": "https://registry.npmjs.org/json-schema-migrate/-/json-schema-migrate-2.0.0.tgz",
-      "integrity": "sha512-r38SVTtojDRp4eD6WsCqiE0eNDt4v1WalBXb9cyZYw9ai5cGtBwzRNWjHzJl38w6TxFkXAIA7h+fyX3tnrAFhQ==",
-      "dev": true,
-      "license": "MIT",
-      "dependencies": {
-        "ajv": "^8.0.0"
-      }
-    },
     "node_modules/json-schema-traverse": {
       "version": "1.0.0",
       "resolved": "https://registry.npmjs.org/json-schema-traverse/-/json-schema-traverse-1.0.0.tgz",
@@ -231,62 +76,6 @@
       "dev": true,
       "license": "MIT"
     },
-    "node_modules/json5": {
-      "version": "2.2.3",
-      "resolved": "https://registry.npmjs.org/json5/-/json5-2.2.3.tgz",
-      "integrity": "sha512-XmOWe7eyHYH14cLdVPoyg+GOH3rYX++KpzrylJwSW98t3Nk+U8XOl8FWKOgwtzdb8lXGf6zYwDUzeHMWfxasyg==",
-      "dev": true,
-      "license": "MIT",
-      "bin": {
-        "json5": "lib/cli.js"
-      },
-      "engines": {
-        "node": ">=6"
-      }
-    },
-    "node_modules/minimatch": {
-      "version": "3.1.5",
-      "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-3.1.5.tgz",
-      "integrity": "sha512-VgjWUsnnT6n+NUk6eZq77zeFdpW2LWDzP6zFGrCbHXiYNul5Dzqk2HHQ5uFH2DNW5Xbp8+jVzaeNt94ssEEl4w==",
-      "dev": true,
-      "license": "ISC",
-      "dependencies": {
-        "brace-expansion": "^1.1.7"
-      },
-      "engines": {
-        "node": "*"
-      }
-    },
-    "node_modules/minimist": {
-      "version": "1.2.8",
-      "resolved": "https://registry.npmjs.org/minimist/-/minimist-1.2.8.tgz",
-      "integrity": "sha512-2yyAR8qBkN3YuheJanUpWC5U3bb5osDywNB8RzDVlDwDHbocAJveqqj1u8+SVD7jkWT4yvsHCpWqqWqAxb0zCA==",
-      "dev": true,
-      "license": "MIT",
-      "funding": {
-        "url": "https://github.com/sponsors/ljharb"
-      }
-    },
-    "node_modules/once": {
-      "version": "1.4.0",
-      "resolved": "https://registry.npmjs.org/once/-/once-1.4.0.tgz",
-      "integrity": "sha512-lNaJgI+2Q5URQBkccEKHTQOPaXdUxnZZElQTZY0MFUAuaEqe1E+Nyvgdz/aIyNi6Z9MzO5dv1H8n58/GELp3+w==",
-      "dev": true,
-      "license": "ISC",
-      "dependencies": {
-        "wrappy": "1"
-      }
-    },
-    "node_modules/path-is-absolute": {
-      "version": "1.0.1",
-      "resolved": "https://registry.npmjs.org/path-is-absolute/-/path-is-absolute-1.0.1.tgz",
-      "integrity": "sha512-AVbw3UJ2e9bq64vSaS9Am0fje1Pa8pbGqTTsmXfaIiMpnr5DlDhfJOuLj9Sf95ZPVDAUerDfEk88MPmPe7UCQg==",
-      "dev": true,
-      "license": "MIT",
-      "engines": {
-        "node": ">=0.10.0"
-      }
-    },
     "node_modules/require-from-string": {
       "version": "2.0.2",
       "resolved": "https://registry.npmjs.org/require-from-string/-/require-from-string-2.0.2.tgz",
@@ -296,20 +85,6 @@
       "engines": {
         "node": ">=0.10.0"
       }
-    },
-    "node_modules/sprintf-js": {
-      "version": "1.0.3",
-      "resolved": "https://registry.npmjs.org/sprintf-js/-/sprintf-js-1.0.3.tgz",
-      "integrity": "sha512-D9cPgkvLlV3t3IzL0D0YLvGA9Ahk4PcvVwUbN0dSGr1aP0Nrt4AEnTUbuGvquEC0mA64Gqt1fzirlRs5ibXx8g==",
-      "dev": true,
-      "license": "BSD-3-Clause"
-    },
-    "node_modules/wrappy": {
-      "version": "1.0.2",
-      "resolved": "https://registry.npmjs.org/wrappy/-/wrappy-1.0.2.tgz",
-      "integrity": "sha512-l4Sp/DRseor9wL6EvV2+TuQn63dMkPjZ/sp9XkghTEbV9KlPS1xUsZ3u7/IQO4wxtcFB4bgpQPRcR3QCvezPcQ==",
-      "dev": true,
-      "license": "ISC"
     }
   }
 }

rules/scripts/package.json

@@ -3,10 +3,7 @@
   "private": true,
   "description": "Validator tooling for rules/ - combine-rules.yml installs from the lockfile so nothing unpinned reaches a job that pushes to main.",
   "devDependencies": {
-    "ajv-cli": "5.0.0",
+    "ajv": "8.20.0",
     "ajv-formats": "3.0.1"
-  },
-  "overrides": {
-    "fast-json-patch": "3.1.1"
   }
 }

rules/scripts/validate.js

@@ -0,0 +1,63 @@
+#!/usr/bin/env node
+/**
+ * Validates every rule file and rules/guardrails.json against its JSON Schema
+ * (draft 2020-12). combine-rules.yml runs this before combine.js; any invalid
+ * file fails the run with exit code 1.
+ *
+ * Run: node rules/scripts/validate.js   (after `npm ci --ignore-scripts` in rules/scripts)
+ */
+
+const fs = require('fs');
+const path = require('path');
+const Ajv2020 = require('ajv/dist/2020');
+const addFormats = require('ajv-formats');
+
+const rulesRoot = path.resolve(__dirname, '..');
+
+const jsonFilesIn = (dir) =>
+  fs.readdirSync(path.join(rulesRoot, dir)).filter(f => f.endsWith('.json')).sort().map(f => `${dir}/${f}`);
+
+const sets = [
+  { schema: 'gather-rule.schema.json', files: jsonFilesIn('gather') },
+  { schema: 'analyze-rule.schema.json', files: jsonFilesIn('analyze') },
+  { schema: 'ime-log-pattern.schema.json', files: jsonFilesIn('ime-log-patterns') },
+  // guardrails.json values are baked verbatim into generated TypeScript that is
+  // auto-committed to main and shipped to the portal - the schema forbids control
+  // and line-terminator characters so a "data" PR can never become code.
+  { schema: 'guardrails.schema.json', files: ['guardrails.json'] }
+];
+
+const ajv = new Ajv2020({ allErrors: true });
+addFormats(ajv);
+
+let invalid = 0;
+
+for (const set of sets) {
+  const validate = ajv.compile(JSON.parse(fs.readFileSync(path.join(rulesRoot, 'schema', set.schema), 'utf8')));
+
+  for (const file of set.files) {
+    let data;
+    try {
+      data = JSON.parse(fs.readFileSync(path.join(rulesRoot, file), 'utf8'));
+    } catch (err) {
+      invalid++;
+      console.error(`INVALID rules/${file}: ${err.message}`);
+      continue;
+    }
+
+    if (!validate(data)) {
+      invalid++;
+      console.error(`INVALID rules/${file}`);
+      for (const e of validate.errors) {
+        console.error(`  ${e.instancePath || '(root)'} ${e.message} ${JSON.stringify(e.params)}`);
+      }
+    }
+  }
+
+  console.log(`${set.schema}: ${set.files.length} files checked`);
+}
+
+if (invalid > 0) {
+  console.error(`ERROR: ${invalid} file(s) failed schema validation`);
+  process.exit(1);
+}

scripts/Bootstrap/Install-AutopilotMonitor.Tests.ps1

@@ -30,6 +30,17 @@ BeforeAll {
         return $dir
     }
 
+    # What Test-FileCatalog -Detailed returns, reduced to the fields the check reads.
+    function New-FakeCatalogResult {
+        param(
+            [string]$Status = 'Valid',
+            [object]$Signature,
+            [hashtable]$CatalogItems = @{ 'AutopilotMonitor.Agent.exe' = 'AA' },
+            [hashtable]$PathItems = @{ 'AutopilotMonitor.Agent.exe' = 'AA' }
+        )
+        return [pscustomobject]@{ Status = $Status; Signature = $Signature; CatalogItems = $CatalogItems; PathItems = $PathItems }
+    }
+
     function New-FakeProfile {
         param([string]$Name, [double]$AgeMinutes)
         $p = Join-Path $TestDrive $Name
@@ -322,86 +333,163 @@ Describe 'logging purity' {
     }
 }
 
-Describe 'Get-UnverifiedAgentBinaries' {
-    # The publisher gate on the extracted payload. What matters is the DECISION: which
-    # payload is accepted, which is refused, and which files are looked at at all.
+Describe 'Get-AgentPayloadRejections' {
+    # The payload catalog check, shared verbatim with the OOBE go-script. What matters is the
+    # DECISION: which payload is accepted and which is refused. Test-FileCatalog is mocked for
+    # the branches; one test runs the real cmdlet over a real (unsigned) catalog.
     BeforeEach {
         Mock Write-Log { }
         $script:ours = 'CN=glueckkanja AG, O=glueckkanja AG, C=DE'
+        $script:dir = New-FakePayload -Files @('AutopilotMonitor.Agent.exe', 'AutopilotMonitor.Agent.cat')
     }
 
-    It 'accepts a payload where every binary of ours is validly signed by us' {
-        $dir = New-FakePayload -Files @('AutopilotMonitor.Agent.exe', 'AutopilotMonitor.Shared.dll')
-        Mock Get-AuthenticodeSignature { New-FakeSignature -Status 'Valid' -Subject $script:ours }
-        Get-UnverifiedAgentBinaries -Path $dir | Should -BeNullOrEmpty
+    It 'accepts a payload whose catalog is signed by us and matches every file' {
+        Mock Test-FileCatalog { New-FakeCatalogResult -Signature (New-FakeSignature -Status 'Valid' -Subject $script:ours) }
+        Get-AgentPayloadRejections -Path $script:dir -ExpectedPublisher $ExpectedPublisher | Should -BeNullOrEmpty
     }
 
-    It 'checks the binaries we author and leaves third-party assemblies alone' {
-        $dir = New-FakePayload -Files @(
-            'AutopilotMonitor.Agent.exe',
-            'AutopilotMonitor.Agent.exe.config',
-            'AutopilotMonitor.Shared.dll',
-            'Newtonsoft.Json.dll',
-            'System.Management.Automation.dll'
-        )
-        $script:checked = @()
-        Mock Get-AuthenticodeSignature {
-            $script:checked += (Split-Path $FilePath -Leaf)
-            New-FakeSignature -Status 'Valid' -Subject $script:ours
+    It 'asks for the detailed result and skips the catalog itself' {
+        Mock Test-FileCatalog {
+            $script:call = @{ Skip = $FilesToSkip; Detailed = [bool]$Detailed; Catalog = $CatalogFilePath; Path = $Path }
+            New-FakeCatalogResult -Signature (New-FakeSignature -Status 'Valid' -Subject $script:ours)
         }
-
-        Get-UnverifiedAgentBinaries -Path $dir | Should -BeNullOrEmpty
-        $script:checked | Should -Be @('AutopilotMonitor.Agent.exe', 'AutopilotMonitor.Shared.dll')
+        Get-AgentPayloadRejections -Path $script:dir -ExpectedPublisher $ExpectedPublisher | Out-Null
+        $script:call.Skip | Should -Be @('AutopilotMonitor.Agent.cat')
+        $script:call.Detailed | Should -BeTrue
+        $script:call.Catalog | Should -Be (Join-Path $script:dir 'AutopilotMonitor.Agent.cat')
+        $script:call.Path | Should -Be $script:dir
     }
 
-    It 'refuses an unsigned agent executable' {
+    It 'refuses a payload without catalog and does not run the check at all' {
         $dir = New-FakePayload -Files @('AutopilotMonitor.Agent.exe')
-        Mock Get-AuthenticodeSignature { New-FakeSignature -Status 'NotSigned' -Subject $null }
-        (Get-UnverifiedAgentBinaries -Path $dir).Count | Should -Be 1
+        Mock Test-FileCatalog { throw 'must not run' }
+        $r = Get-AgentPayloadRejections -Path $dir -ExpectedPublisher $ExpectedPublisher
+        $r.Count | Should -Be 1
+        $r[0] | Should -BeLike '*AutopilotMonitor.Agent.cat missing*'
+        Should -Invoke Test-FileCatalog -Times 0
     }
 
-    It 'refuses a tampered binary (hash mismatch)' {
-        $dir = New-FakePayload -Files @('AutopilotMonitor.Agent.exe')
-        Mock Get-AuthenticodeSignature { New-FakeSignature -Status 'HashMismatch' -Subject $script:ours }
-        (Get-UnverifiedAgentBinaries -Path $dir).Count | Should -Be 1
+    It 'refuses matching files under an unsigned catalog (Status alone is not the verdict)' {
+        Mock Test-FileCatalog { New-FakeCatalogResult -Status 'Valid' -Signature (New-FakeSignature -Status 'NotSigned' -Subject $null) }
+        $r = Get-AgentPayloadRejections -Path $script:dir -ExpectedPublisher $ExpectedPublisher
+        $r.Count | Should -Be 1
+        $r[0] | Should -BeLike '*signature status NotSigned*'
     }
 
-    It 'refuses a signature whose chain does not validate' {
-        $dir = New-FakePayload -Files @('AutopilotMonitor.Agent.exe')
-        Mock Get-AuthenticodeSignature { New-FakeSignature -Status 'UnknownError' -Subject $script:ours }
-        (Get-UnverifiedAgentBinaries -Path $dir).Count | Should -Be 1
+    It 'refuses a catalog whose signature does not validate' {
+        Mock Test-FileCatalog { New-FakeCatalogResult -Signature (New-FakeSignature -Status 'HashMismatch' -Subject $script:ours) }
+        (Get-AgentPayloadRejections -Path $script:dir -ExpectedPublisher $ExpectedPublisher).Count | Should -Be 1
     }
 
-    It 'refuses a validly signed binary from a different publisher' {
-        $dir = New-FakePayload -Files @('AutopilotMonitor.Agent.exe')
-        Mock Get-AuthenticodeSignature { New-FakeSignature -Status 'Valid' -Subject 'CN=Somebody Else, O=Somebody Else Ltd, C=US' }
-        (Get-UnverifiedAgentBinaries -Path $dir).Count | Should -Be 1
+    It 'refuses a validly signed catalog from another publisher' {
+        Mock Test-FileCatalog { New-FakeCatalogResult -Signature (New-FakeSignature -Status 'Valid' -Subject 'CN=Somebody Else, O=Somebody Else Ltd, C=US') }
+        $r = Get-AgentPayloadRejections -Path $script:dir -ExpectedPublisher $ExpectedPublisher
+        $r.Count | Should -Be 1
+        $r[0] | Should -BeLike '*does not match*'
     }
 
-    It 'refuses the whole payload when a single DLL of ours is swapped' {
-        $dir = New-FakePayload -Files @('AutopilotMonitor.Agent.exe', 'AutopilotMonitor.Shared.dll')
-        Mock Get-AuthenticodeSignature {
-            if ((Split-Path $FilePath -Leaf) -eq 'AutopilotMonitor.Shared.dll') {
-                New-FakeSignature -Status 'Valid' -Subject 'CN=Somebody Else, O=Somebody Else Ltd, C=US'
-            } else {
-                New-FakeSignature -Status 'Valid' -Subject $script:ours
-            }
+    It 'names the missing, extra and changed files of a payload that differs' {
+        Mock Test-FileCatalog {
+            New-FakeCatalogResult -Status 'ValidationFailed' -Signature (New-FakeSignature -Status 'Valid' -Subject $script:ours) `
+                -CatalogItems @{ 'a.dll' = 'AA'; 'b.dll' = 'BB'; 'c.config' = 'CC' } `
+                -PathItems @{ 'a.dll' = 'AA'; 'b.dll' = 'XX'; 'evil.dll' = 'EE' }
         }
-        $rejected = Get-UnverifiedAgentBinaries -Path $dir
-        $rejected.Count | Should -Be 1
-        $rejected[0] | Should -BeLike 'AutopilotMonitor.Shared.dll*'
+        $r = Get-AgentPayloadRejections -Path $script:dir -ExpectedPublisher $ExpectedPublisher
+        $r.Count | Should -Be 1
+        $r[0] | Should -BeLike '*missing: c.config;*'
+        $r[0] | Should -BeLike '*extra: evil.dll;*'
+        $r[0] | Should -BeLike '*changed: b.dll)*'
     }
 
-    It 'refuses when the signature check itself throws' {
-        $dir = New-FakePayload -Files @('AutopilotMonitor.Agent.exe')
-        Mock Get-AuthenticodeSignature { throw 'access denied' }
-        (Get-UnverifiedAgentBinaries -Path $dir).Count | Should -Be 1
+    It 'reports a bad signature and a differing payload together' {
+        Mock Test-FileCatalog {
+            New-FakeCatalogResult -Status 'ValidationFailed' -Signature (New-FakeSignature -Status 'NotSigned' -Subject $null) `
+                -PathItems @{ 'AutopilotMonitor.Agent.exe' = 'AA'; 'extra.dll' = 'EE' }
+        }
+        (Get-AgentPayloadRejections -Path $script:dir -ExpectedPublisher $ExpectedPublisher).Count | Should -Be 2
     }
 
-    It 'refuses an empty payload instead of passing it' {
-        $dir = New-FakePayload -Files @('Newtonsoft.Json.dll')
-        Mock Get-AuthenticodeSignature { New-FakeSignature -Status 'Valid' -Subject $script:ours }
-        (Get-UnverifiedAgentBinaries -Path $dir).Count | Should -Be 1
+    It 'refuses when the catalog check itself throws' {
+        Mock Test-FileCatalog { throw 'access denied' }
+        $r = Get-AgentPayloadRejections -Path $script:dir -ExpectedPublisher $ExpectedPublisher
+        $r.Count | Should -Be 1
+        $r[0] | Should -BeLike '*failed to run*access denied*'
+    }
+
+    It 'refuses when the catalog check returns nothing' {
+        Mock Test-FileCatalog { }
+        $r = Get-AgentPayloadRejections -Path $script:dir -ExpectedPublisher $ExpectedPublisher
+        $r[0] | Should -BeLike '*returned nothing*'
+    }
+
+    It 'returns a single rejection as an array' {
+        Mock Test-FileCatalog { throw 'boom' }
+        $r = Get-AgentPayloadRejections -Path $script:dir -ExpectedPublisher $ExpectedPublisher
+        ($r -is [array]) | Should -BeTrue
+    }
+
+    It 'catches an extra file with the real cmdlet over a real catalog' {
+        # Real New-FileCatalog/Test-FileCatalog over text files: no PE, no process, no mock.
+        $dir = New-FakePayload -Files @('a.txt', 'b.txt')
+        $catalog = Join-Path $TestDrive "real-$([guid]::NewGuid()).cat"
+        New-FileCatalog -Path $dir -CatalogFilePath $catalog -CatalogVersion 2.0 | Out-Null
+        Copy-Item $catalog (Join-Path $dir 'AutopilotMonitor.Agent.cat')
+        Set-Content -Path (Join-Path $dir 'extra.txt') -Value 'planted'
+
+        $r = Get-AgentPayloadRejections -Path $dir -ExpectedPublisher $ExpectedPublisher
+
+        ($r -join ' | ') | Should -BeLike '*extra: extra.txt;*'
+        ($r -join ' | ') | Should -BeLike '*signature status NotSigned*'
+    }
+}
+
+Describe 'Install-VerifiedAgentPayload' {
+    BeforeEach {
+        Mock Write-Log { }
+        $src = New-FakePayload -Files @('AutopilotMonitor.Agent.exe', 'AutopilotMonitor.Agent.cat', 'Newtonsoft.Json.dll')
+        $script:zip = Join-Path $TestDrive "payload-$([guid]::NewGuid()).zip"
+        Compress-Archive -Path (Join-Path $src '*') -DestinationPath $script:zip
+        $script:staging = Join-Path $TestDrive "staging-$([guid]::NewGuid())"
+        $script:target = Join-Path $TestDrive "agent-$([guid]::NewGuid())"
+    }
+
+    It 'moves an accepted payload into the agent folder and removes the staging folder' {
+        Mock Get-AgentPayloadRejections { , @() }
+        Install-VerifiedAgentPayload -ZipPath $script:zip -StagingPath $script:staging -TargetPath $script:target -ExpectedPublisher $ExpectedPublisher
+        @(Get-ChildItem $script:target | ForEach-Object Name | Sort-Object) |
+            Should -Be @('AutopilotMonitor.Agent.cat', 'AutopilotMonitor.Agent.exe', 'Newtonsoft.Json.dll')
+        Test-Path $script:staging | Should -BeFalse
+    }
+
+    It 'verifies the staging folder, never the agent folder' {
+        Mock Get-AgentPayloadRejections { $script:checked = $Path; , @() }
+        Install-VerifiedAgentPayload -ZipPath $script:zip -StagingPath $script:staging -TargetPath $script:target -ExpectedPublisher $ExpectedPublisher
+        $script:checked | Should -Be $script:staging
+    }
+
+    It 'leaves the agent folder untouched when the payload is refused, and cleans up' {
+        Mock Get-AgentPayloadRejections { , @('payload differs from catalog (missing: none; extra: evil.dll; changed: none)') }
+        { Install-VerifiedAgentPayload -ZipPath $script:zip -StagingPath $script:staging -TargetPath $script:target -ExpectedPublisher $ExpectedPublisher } |
+            Should -Throw '*Payload verification FAILED*evil.dll*'
+        Test-Path $script:target | Should -BeFalse
+        Test-Path $script:staging | Should -BeFalse
+    }
+
+    It 'removes leftovers from the agent folder before the payload moves in' {
+        New-Item -ItemType Directory -Path $script:target | Out-Null
+        Set-Content -Path (Join-Path $script:target 'stale.dll') -Value 'x'
+        Mock Get-AgentPayloadRejections { , @() }
+        Install-VerifiedAgentPayload -ZipPath $script:zip -StagingPath $script:staging -TargetPath $script:target -ExpectedPublisher $ExpectedPublisher
+        Test-Path (Join-Path $script:target 'stale.dll') | Should -BeFalse
+        Test-Path (Join-Path $script:target 'AutopilotMonitor.Agent.exe') | Should -BeTrue
+    }
+
+    It 'starts from an empty staging folder even when a previous run left one behind' {
+        New-Item -ItemType Directory -Path $script:staging | Out-Null
+        Set-Content -Path (Join-Path $script:staging 'leftover.dll') -Value 'x'
+        Mock Get-AgentPayloadRejections { $script:seen = @(Get-ChildItem $Path | ForEach-Object Name); , @() }
+        Install-VerifiedAgentPayload -ZipPath $script:zip -StagingPath $script:staging -TargetPath $script:target -ExpectedPublisher $ExpectedPublisher
+        $script:seen | Should -Not -Contain 'leftover.dll'
     }
 }
 
@@ -410,4 +498,8 @@ Describe 'publisher pin' {
         $ExpectedPublisher | Should -Be '*O=glueckkanja AG*'
         $ExpectedPublisher | Should -Not -Match '^[0-9A-Fa-f]{40}$'
     }
+
+    It 'names the payload catalog the release pipeline signs' {
+        $AgentCatalogName | Should -Be 'AutopilotMonitor.Agent.cat'
+    }
 }

scripts/Bootstrap/Install-AutopilotMonitor.ps1

@@ -11,10 +11,12 @@
          freshly provisioned Windows 365 Cloud PC.
       2. Downloads the monitoring agent ZIP from Azure Blob Storage
       3. Verifies integrity via SHA-256 hash from the version manifest (mandatory)
-      4. Extracts the agent into %ProgramData%\AutopilotMonitor\Agent
-      5. Verifies the Authenticode signature and publisher of every binary we author
-      6. Runs the agent in --install mode (registers Scheduled Task, spawns the runtime)
-      7. Verifies the runtime process actually launched
+      4. Extracts the agent into a fresh staging folder
+      5. Verifies the payload against its signed catalog (AutopilotMonitor.Agent.cat):
+         signed by the expected publisher, every file present and unchanged, nothing extra
+      6. Moves the verified payload into %ProgramData%\AutopilotMonitor\Agent
+      7. Runs the agent in --install mode (registers Scheduled Task, spawns the runtime)
+      8. Verifies the runtime process actually launched
     Agent self-destructs when enrollment completes.
 
     Relax flow (guards 2+3): a real user profile (or a real LastLoggedOnUser)
@@ -77,6 +79,14 @@
       PowerShell 5.1 (IME) reads scripts without BOM as ANSI, corrupting multi-byte chars.
 
 .CHANGELOG
+    2026-10-13  v2.6  Payload catalog: the agent ZIP carries AutopilotMonitor.Agent.cat,
+                      a catalog of every payload file signed by the expected publisher.
+                      The ZIP is extracted into a staging folder and moved into the
+                      agent folder only when Test-FileCatalog finds the catalog signed
+                      by the expected publisher and every file present, unchanged and
+                      nothing extra. A rejected payload is deleted, so the next run
+                      starts clean. Replaces the v2.5 gate, which covered only the
+                      AutopilotMonitor binaries.
     2026-09-08  v2.5  Publisher gate: every AutopilotMonitor binary in the extracted
                       payload must carry a valid Authenticode signature from the
                       expected publisher before --install runs. The SHA-256 manifest
@@ -134,16 +144,20 @@ param(
 )
 
 # Script version (bump on meaningful changes; see .CHANGELOG above)
-$ScriptVersion = "2.5"
+$ScriptVersion = "2.6"
 
-# The publisher every binary we author must carry. A -like pattern on the subject, NOT a
+# The publisher the payload catalog must carry. A -like pattern on the subject, NOT a
 # thumbprint: certificates are renewed, deployed copies of this script are not. Same value
 # the loader pins for this script.
 $ExpectedPublisher = "*O=glueckkanja AG*"
 
+# Signed catalog at the root of every agent ZIP: the SHA-256 of every other payload file.
+$AgentCatalogName = "AutopilotMonitor.Agent.cat"
+
 # Configuration - Everything in ProgramData for easy cleanup
 $AgentBasePath = "$env:ProgramData\AutopilotMonitor"
 $AgentBinPath = "$AgentBasePath\Agent"
+$AgentStagingPath = "$AgentBasePath\Agent-Install"
 $AgentLogPath = "$AgentBasePath\Logs"
 $LogFile = "$AgentLogPath\bootstrap-script.log"
 
@@ -340,56 +354,107 @@ function Get-BootstrapDecision {
     return [pscustomobject]@{ Install = $true; ReasonCode = $null; RelaxActive = $relax.Active }
 }
 
-# Returns the list of reasons why the extracted payload is not ours; an empty result means
-# every binary we author carries a valid Authenticode signature from the expected publisher.
-# Third-party assemblies keep their own publisher's signature and are not checked here.
-# Any error is a rejection: an unreadable signature is an unverified binary.
-function Get-UnverifiedAgentBinaries {
-    param([string]$Path)
-
-    $rejected = @()
-    $ourBinaries = @(
-        Get-ChildItem -Path $Path -Filter 'AutopilotMonitor.*' -File -ErrorAction SilentlyContinue |
-            Where-Object { $_.Extension -in @('.exe', '.dll') }
+### BEGIN shared payload verification -- embedded verbatim into the OOBE go-script by
+### OobeBootstrapScriptGenerator (D-354). Keep it self-contained: parameters only, no
+### script-scope variables, no Write-Log.
+# Returns the reasons why the extracted payload in $Path must not run; an empty result means
+# it matches its signed catalog exactly. Test-FileCatalog's Status covers the hashes only (an
+# unsigned catalog over matching files reports Valid), so the signature and its publisher are
+# judged on their own. Any error is a rejection: a payload that cannot be verified is an
+# unverified payload. The comma keeps a single rejection an ARRAY: PowerShell unrolls a
+# one-element result to a bare string, and the caller reads .Count.
+function Get-AgentPayloadRejections {
+    param(
+        [string]$Path,
+        [string]$ExpectedPublisher,
+        [string]$CatalogName = 'AutopilotMonitor.Agent.cat'
     )
 
-    # An empty set is a rejection, not a pass: it means the extraction produced nothing we
-    # recognise, and the alternative would be a gate that silently approves an empty folder.
-    # The comma keeps a single rejection an ARRAY: PowerShell unrolls a one-element result to
-    # a bare string, and the caller reads .Count and indexes into it.
-    if ($ourBinaries.Count -eq 0) {
-        return , @("no AutopilotMonitor binaries found in $Path")
+    $catalog = Join-Path $Path $CatalogName
+    if (-not (Test-Path -LiteralPath $catalog -PathType Leaf)) {
+        return , @("catalog $CatalogName missing")
     }
 
-    foreach ($binary in $ourBinaries) {
-        try {
-            $sig = Get-AuthenticodeSignature -FilePath $binary.FullName -ErrorAction Stop
-        }
-        catch {
-            $rejected += "$($binary.Name): signature check failed to run ($($_.Exception.Message))"
-            continue
-        }
+    try {
+        $result = Test-FileCatalog -Path $Path -CatalogFilePath $catalog -FilesToSkip $CatalogName -Detailed -ErrorAction Stop
+    }
+    catch {
+        return , @("catalog check failed to run ($($_.Exception.Message))")
+    }
+    if ($null -eq $result) {
+        return , @('catalog check returned nothing')
+    }
 
-        if ($null -eq $sig) {
-            $rejected += "$($binary.Name): signature check returned nothing"
-            continue
-        }
+    $rejected = @()
+    $signature = $result.Signature
+    if ($null -eq $signature -or "$($signature.Status)" -ne 'Valid') {
+        $status = if ($null -eq $signature) { 'missing' } else { "$($signature.Status)" }
+        $rejected += "catalog signature status $status"
+    }
+    elseif ($null -eq $signature.SignerCertificate -or $signature.SignerCertificate.Subject -notlike $ExpectedPublisher) {
+        $subject = if ($null -eq $signature.SignerCertificate) { '' } else { $signature.SignerCertificate.Subject }
+        $rejected += "catalog publisher '$subject' does not match $ExpectedPublisher"
+    }
 
-        $subject = if ($sig.SignerCertificate) { $sig.SignerCertificate.Subject } else { '<unsigned>' }
-
-        if ($sig.Status -ne 'Valid') {
-            $rejected += "$($binary.Name): signature status $($sig.Status)"
-            continue
-        }
-        if ($subject -notlike $ExpectedPublisher) {
-            $rejected += "$($binary.Name): unexpected publisher $subject"
-            continue
+    if ("$($result.Status)" -ne 'Valid') {
+        $expected = if ($null -eq $result.CatalogItems) { @{} } else { $result.CatalogItems }
+        $actual = if ($null -eq $result.PathItems) { @{} } else { $result.PathItems }
+        $names = {
+            param($list)
+            if ($list.Count -eq 0) { return 'none' }
+            $shown = ($list | Select-Object -First 5) -join ', '
+            if ($list.Count -gt 5) { $shown += " (+$($list.Count - 5) more)" }
+            return $shown
         }
+        $missing = @($expected.Keys | Where-Object { -not $actual.ContainsKey($_) } | Sort-Object)
+        $extra = @($actual.Keys | Where-Object { -not $expected.ContainsKey($_) } | Sort-Object)
+        $changed = @($expected.Keys | Where-Object { $actual.ContainsKey($_) -and $actual[$_] -ne $expected[$_] } | Sort-Object)
+        $rejected += "payload differs from catalog (missing: $(& $names $missing); extra: $(& $names $extra); changed: $(& $names $changed))"
     }
 
     return , $rejected
 }
 
+# Extracts the agent ZIP into a fresh staging folder, verifies it there and moves it into the
+# agent folder only when it matches its signed catalog. A rejected payload is deleted with the
+# staging folder: nothing unverified lands where SYSTEM runs the agent, and the next run starts
+# clean instead of finding a half-installed agent. The caller has made sure no agent is
+# installed, so anything already in the agent folder is a leftover and goes first.
+function Install-VerifiedAgentPayload {
+    param(
+        [string]$ZipPath,
+        [string]$StagingPath,
+        [string]$TargetPath,
+        [string]$ExpectedPublisher,
+        [string]$CatalogName = 'AutopilotMonitor.Agent.cat'
+    )
+
+    if (Test-Path -LiteralPath $StagingPath) {
+        Remove-Item -LiteralPath $StagingPath -Recurse -Force -ErrorAction Stop
+    }
+    New-Item -Path $StagingPath -ItemType Directory -Force | Out-Null
+    try {
+        Expand-Archive -Path $ZipPath -DestinationPath $StagingPath -Force -ErrorAction Stop
+
+        $rejections = Get-AgentPayloadRejections -Path $StagingPath -ExpectedPublisher $ExpectedPublisher -CatalogName $CatalogName
+        if ($rejections.Count -gt 0) {
+            throw "Payload verification FAILED ($($rejections -join '; ')). Refusing to install an agent that is not the published one."
+        }
+
+        if (Test-Path -LiteralPath $TargetPath) {
+            Get-ChildItem -LiteralPath $TargetPath -Force | Remove-Item -Recurse -Force -ErrorAction Stop
+        }
+        else {
+            New-Item -Path $TargetPath -ItemType Directory -Force | Out-Null
+        }
+        Get-ChildItem -LiteralPath $StagingPath -Force | Move-Item -Destination $TargetPath -Force -ErrorAction Stop
+    }
+    finally {
+        Remove-Item -LiteralPath $StagingPath -Recurse -Force -ErrorAction SilentlyContinue
+    }
+}
+### END shared payload verification
+
 # Main bootstrap flow. Only invoked when the script is run directly (see entry guard
 # at the bottom); never on dot-source.
 function Invoke-Bootstrap {
@@ -408,117 +473,106 @@ function Invoke-Bootstrap {
 
         Write-Log "Pre-flight checks passed -- no prior deployment, no productive-device signals, within bootstrap window (relaxActive=$($decision.RelaxActive))"
 
-        # Download and extract agent binaries
+        # Download, verify and install the agent binaries. Guard 5 has already returned when an
+        # agent exe exists, so this always starts without an installed agent.
         $agentExePath = Join-Path $AgentBinPath "AutopilotMonitor.Agent.exe"
+        Write-Log "Downloading agent from $AgentDownloadUrl..."
 
-        if (Test-Path $agentExePath) {
-            Write-Log "Agent already installed at $agentExePath"
-        }
-        else {
-            Write-Log "Downloading agent from $AgentDownloadUrl..."
+        try {
+            # Derive manifest URL from the agent download URL (same blob container)
+            $versionJsonUrl = $AgentDownloadUrl -replace '[^/]+$', $VersionJsonName
+            $expectedSha256 = $null
+            $manifestAttempts = 3
+            $manifestDelays = @(1, 2)   # short backoff between attempts; do not block enrollment
 
-            try {
-                # Derive manifest URL from the agent download URL (same blob container)
-                $versionJsonUrl = $AgentDownloadUrl -replace '[^/]+$', $VersionJsonName
-                $expectedSha256 = $null
-                $manifestAttempts = 3
-                $manifestDelays = @(1, 2)   # short backoff between attempts; do not block enrollment
-
-                for ($attempt = 1; $attempt -le $manifestAttempts; $attempt++) {
-                    try {
-                        Write-Log "Fetching $VersionJsonName (attempt ${attempt}/${manifestAttempts}) from $versionJsonUrl for integrity verification..."
-                        $versionJsonResponse = Invoke-RestMethod -Uri $versionJsonUrl -UseBasicParsing -TimeoutSec 5 -ErrorAction Stop
-                        if ($versionJsonResponse.sha256) {
-                            $expectedSha256 = $versionJsonResponse.sha256.ToLowerInvariant()
-                            Write-Log "SHA-256 hash from manifest: $expectedSha256 (version: $($versionJsonResponse.version))"
-                        } else {
-                            Write-Log "WARNING: Manifest has no sha256 field - integrity cannot be verified."
-                        }
-                        break   # manifest reached (with or without hash); retrying would not help
+            for ($attempt = 1; $attempt -le $manifestAttempts; $attempt++) {
+                try {
+                    Write-Log "Fetching $VersionJsonName (attempt ${attempt}/${manifestAttempts}) from $versionJsonUrl for integrity verification..."
+                    $versionJsonResponse = Invoke-RestMethod -Uri $versionJsonUrl -UseBasicParsing -TimeoutSec 5 -ErrorAction Stop
+                    if ($versionJsonResponse.sha256) {
+                        $expectedSha256 = $versionJsonResponse.sha256.ToLowerInvariant()
+                        Write-Log "SHA-256 hash from manifest: $expectedSha256 (version: $($versionJsonResponse.version))"
+                    } else {
+                        Write-Log "WARNING: Manifest has no sha256 field - integrity cannot be verified."
                     }
-                    catch {
-                        Write-Log "WARNING: Could not fetch integrity manifest (attempt $attempt): $($_.Exception.Message)"
-                        if ($attempt -lt $manifestAttempts) {
-                            Start-Sleep -Seconds $manifestDelays[$attempt - 1]
-                        }
+                    break   # manifest reached (with or without hash); retrying would not help
+                }
+                catch {
+                    Write-Log "WARNING: Could not fetch integrity manifest (attempt $attempt): $($_.Exception.Message)"
+                    if ($attempt -lt $manifestAttempts) {
+                        Start-Sleep -Seconds $manifestDelays[$attempt - 1]
                     }
                 }
-
-                $zipPath = Join-Path $env:TEMP "AutopilotMonitor-Agent.zip"
-                $maxDownloadAttempts = 3
-                $downloadAttempt = 0
-
-                do {
-                    $downloadAttempt++
-                    try {
-                        Write-Log "Download attempt ${downloadAttempt}/${maxDownloadAttempts}"
-                        $downloadStopwatch = [System.Diagnostics.Stopwatch]::StartNew()
-                        Invoke-WebRequest `
-                            -Uri $AgentDownloadUrl `
-                            -OutFile $zipPath `
-                            -UseBasicParsing `
-                            -TimeoutSec 30 `
-                            -ErrorAction Stop
-                        $downloadStopwatch.Stop()
-                        $downloadSeconds = [math]::Round($downloadStopwatch.Elapsed.TotalSeconds, 1)
-                        Write-Log "Downloaded agent to $zipPath (took ${downloadSeconds}s)"
-                        break
-                    }
-                    catch {
-                        if ($downloadAttempt -ge $maxDownloadAttempts) {
-                            throw
-                        }
-
-                        $retryDelaysInSeconds = @(2, 4, 8)
-                        $retryDelaySeconds = $retryDelaysInSeconds[$downloadAttempt - 1]
-                        Write-Log "Download failed (attempt $downloadAttempt): $($_.Exception.Message). Retrying in ${retryDelaySeconds}s..."
-                        Start-Sleep -Seconds $retryDelaySeconds
-                    }
-                } while ($downloadAttempt -lt $maxDownloadAttempts)
-
-                # SHA-256 from the version manifest is the only accepted integrity proof.
-                # No hash -> refuse to install an unverified ZIP (SKIP-safe security posture).
-                if (-not $expectedSha256) {
-                    throw "No SHA-256 hash available from manifest - refusing to install unverified agent ZIP."
-                }
-                $actualSha256 = (Get-FileHash $zipPath -Algorithm SHA256).Hash.ToLowerInvariant()
-                Write-Log "Validating SHA-256 hash against manifest"
-                if ($actualSha256 -ne $expectedSha256) {
-                    throw "SHA-256 integrity check FAILED. Expected='$expectedSha256', Actual='$actualSha256'. Download may be tampered or corrupted."
-                }
-                Write-Log "SHA-256 integrity check passed"
-
-                Expand-Archive -Path $zipPath -DestinationPath $AgentBinPath -Force
-                Write-Log "Extracted agent to $AgentBinPath"
-
-                Remove-Item $zipPath -Force -ErrorAction SilentlyContinue
-                Write-Log "Cleaned up temporary files"
-
-                if (-not (Test-Path $agentExePath)) {
-                    throw "Agent executable not found after extraction at $agentExePath"
-                }
-
-                Write-Log "Agent installation completed successfully"
             }
-            catch {
-                Write-Log "ERROR downloading/extracting agent: $($_.Exception.Message)"
-                throw
-            }
-        }
 
-        # Publisher gate. The SHA-256 above proves the ZIP is the one the manifest names, but
-        # that manifest lives in the same container as the ZIP and cannot outlive a compromise
-        # of the download host. The expected publisher travels inside THIS script, which the
-        # loader verified before running it, so the chain from the file an administrator
-        # assigns to the process that starts here stays publisher-anchored end to end.
-        # Fail-closed, exactly like the hash check: running unverified code as SYSTEM is the
-        # risk being removed.
-        Write-Log "Verifying Authenticode signatures of the agent binaries..."
-        $unverified = Get-UnverifiedAgentBinaries -Path $AgentBinPath
-        if ($unverified.Count -gt 0) {
-            throw "Authenticode verification FAILED ($($unverified -join '; ')). Refusing to run an agent that is not signed by the expected publisher."
+            $zipPath = Join-Path $env:TEMP "AutopilotMonitor-Agent.zip"
+            $maxDownloadAttempts = 3
+            $downloadAttempt = 0
+
+            do {
+                $downloadAttempt++
+                try {
+                    Write-Log "Download attempt ${downloadAttempt}/${maxDownloadAttempts}"
+                    $downloadStopwatch = [System.Diagnostics.Stopwatch]::StartNew()
+                    Invoke-WebRequest `
+                        -Uri $AgentDownloadUrl `
+                        -OutFile $zipPath `
+                        -UseBasicParsing `
+                        -TimeoutSec 30 `
+                        -ErrorAction Stop
+                    $downloadStopwatch.Stop()
+                    $downloadSeconds = [math]::Round($downloadStopwatch.Elapsed.TotalSeconds, 1)
+                    Write-Log "Downloaded agent to $zipPath (took ${downloadSeconds}s)"
+                    break
+                }
+                catch {
+                    if ($downloadAttempt -ge $maxDownloadAttempts) {
+                        throw
+                    }
+
+                    $retryDelaysInSeconds = @(2, 4, 8)
+                    $retryDelaySeconds = $retryDelaysInSeconds[$downloadAttempt - 1]
+                    Write-Log "Download failed (attempt $downloadAttempt): $($_.Exception.Message). Retrying in ${retryDelaySeconds}s..."
+                    Start-Sleep -Seconds $retryDelaySeconds
+                }
+            } while ($downloadAttempt -lt $maxDownloadAttempts)
+
+            # SHA-256 from the version manifest is the only accepted integrity proof.
+            # No hash -> refuse to install an unverified ZIP (SKIP-safe security posture).
+            if (-not $expectedSha256) {
+                throw "No SHA-256 hash available from manifest - refusing to install unverified agent ZIP."
+            }
+            $actualSha256 = (Get-FileHash $zipPath -Algorithm SHA256).Hash.ToLowerInvariant()
+            Write-Log "Validating SHA-256 hash against manifest"
+            if ($actualSha256 -ne $expectedSha256) {
+                throw "SHA-256 integrity check FAILED. Expected='$expectedSha256', Actual='$actualSha256'. Download may be tampered or corrupted."
+            }
+            Write-Log "SHA-256 integrity check passed"
+
+            # Payload catalog. The SHA-256 above proves the ZIP is the one the manifest names, but
+            # that manifest lives in the same container as the ZIP and cannot outlive a compromise
+            # of the download host. The catalog is signed by the expected publisher, whose pin
+            # travels inside THIS script -- verified by the loader before it ran -- so the chain
+            # from the file an administrator assigns to every file the agent runs from stays
+            # publisher-anchored end to end. Fail-closed like the hash check: running unverified
+            # code as SYSTEM is the risk being removed.
+            Write-Log "Verifying the payload against its signed catalog ($AgentCatalogName, publisher pin: $ExpectedPublisher)..."
+            Install-VerifiedAgentPayload -ZipPath $zipPath -StagingPath $AgentStagingPath -TargetPath $AgentBinPath -ExpectedPublisher $ExpectedPublisher -CatalogName $AgentCatalogName
+            Write-Log "Payload verified against its signed catalog and installed to $AgentBinPath"
+
+            Remove-Item $zipPath -Force -ErrorAction SilentlyContinue
+            Write-Log "Cleaned up temporary files"
+
+            if (-not (Test-Path $agentExePath)) {
+                throw "Agent executable not found after extraction at $agentExePath"
+            }
+
+            Write-Log "Agent installation completed successfully"
+        }
+        catch {
+            Write-Log "ERROR downloading/verifying agent: $($_.Exception.Message)"
+            throw
         }
-        Write-Log "Authenticode verification passed (publisher pin: $ExpectedPublisher)"
 
         Write-Log "Calling agent install mode (--install)..."
         & $agentExePath --install

src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/Configuration/RemoteConfigMergerCompletenessTests.cs

@@ -50,6 +50,7 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Configuration
             nameof(AgentConfigResponse.MaxBatchSize),
             nameof(AgentConfigResponse.MaxAuthFailures),
             nameof(AgentConfigResponse.AuthFailureTimeoutMinutes),
+            nameof(AgentConfigResponse.InterceptedRegistrationRetryBoots),
             // Merge reads two Collectors knobs (AgentMaxLifetimeMinutes, HelloWaitTimeoutSeconds);
             // the remaining CollectorConfiguration flags flow untouched to DefaultComponentFactory.
             nameof(AgentConfigResponse.Collectors),

src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/Configuration/RemoteConfigMergerTests.cs

@@ -55,6 +55,7 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Configuration
             UnrestrictedMode = true,
             MaxAuthFailures = 7,
             AuthFailureTimeoutMinutes = 15,
+            InterceptedRegistrationRetryBoots = 2,
             EnableImeMatchLog = true,
             Collectors = new CollectorConfiguration
             {
@@ -98,6 +99,7 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Configuration
             Assert.Equal(45, agent.HelloWaitTimeoutSeconds);
             Assert.Equal(7, agent.MaxAuthFailures);
             Assert.Equal(15, agent.AuthFailureTimeoutMinutes);
+            Assert.Equal(2, agent.InterceptedRegistrationRetryBoots);
 
             Assert.NotNull(result);
             Assert.True(result.UnrestrictedModeChanged);
@@ -468,6 +470,50 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Configuration
             Assert.Equal(0, agent.AgentMaxLifetimeMinutes);
         }
 
+        // ============================================================= Intercepted-registration buffer
+
+        [Theory]
+        [InlineData(0)] // clean up at the first intercepted registration
+        [InlineData(3)]
+        public void Merge_copies_intercepted_registration_retry_boots(int remoteValue)
+        {
+            var agent = NewAgentConfig();
+            var remote = FullRemote();
+            remote.InterceptedRegistrationRetryBoots = remoteValue;
+
+            RemoteConfigMerger.Merge(agent, remote);
+
+            Assert.Equal(remoteValue, agent.InterceptedRegistrationRetryBoots);
+        }
+
+        [Fact]
+        public void Merge_keeps_the_agent_default_when_the_backend_sends_no_intercepted_registration_retry_boots()
+        {
+            var agent = NewAgentConfig();
+            var remote = FullRemote();
+            remote.InterceptedRegistrationRetryBoots = null; // backend that predates the field
+
+            RemoteConfigMerger.Merge(agent, remote);
+
+            Assert.Equal(1, agent.InterceptedRegistrationRetryBoots);
+        }
+
+        [Theory]
+        [InlineData(-1, 0)]
+        [InlineData(int.MinValue, 0)]
+        [InlineData(6, 5)]
+        [InlineData(int.MaxValue, 5)]
+        public void Merge_clamps_intercepted_registration_retry_boots_to_zero_through_five(int remoteValue, int expected)
+        {
+            var agent = NewAgentConfig();
+            var remote = FullRemote();
+            remote.InterceptedRegistrationRetryBoots = remoteValue;
+
+            RemoteConfigMerger.Merge(agent, remote);
+
+            Assert.Equal(expected, agent.InterceptedRegistrationRetryBoots);
+        }
+
         // ============================================================= Guard
 
         [Fact]

src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/Configuration/RemoteConfigServiceCacheHardeningTests.cs

@@ -214,6 +214,23 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Configuration
             Assert.Null(loaded.LatestAgentSha256);
         }
 
+        [Theory]
+        [InlineData(0)]
+        [InlineData(3)]
+        [InlineData(null)] // backend that predates the field — the agent keeps its default
+        public void RoundTrip_preserves_intercepted_registration_retry_boots(int? value)
+        {
+            // An intercepted start gets no live config, so the registration buffer it decides on
+            // is the cached value: it must survive the write + read strip untouched.
+            var probe = NewProbe();
+
+            probe.DoCache(new AgentConfigResponse { InterceptedRegistrationRetryBoots = value });
+            var loaded = probe.DoLoad();
+
+            Assert.NotNull(loaded);
+            Assert.Equal(value, loaded.InterceptedRegistrationRetryBoots);
+        }
+
         // ── Helpers ─────────────────────────────────────────────────────────
 
         private CacheProbe NewProbe()

src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/Configuration/RemoteConfigServiceRetryTests.cs

@@ -7,6 +7,7 @@ using AutopilotMonitor.Agent.V2.Core.Logging;
 using AutopilotMonitor.Agent.V2.Core.Monitoring.Transport;
 using AutopilotMonitor.Agent.V2.Core.Security;
 using AutopilotMonitor.Agent.V2.Core.Tests.Harness;
+using AutopilotMonitor.Agent.V2.Core.Tests.Transport;
 using AutopilotMonitor.Shared.Models;
 using Xunit;
 
@@ -101,6 +102,28 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Configuration
             Assert.Equal(401, svc.LastFetchAuthStatusCode);
         }
 
+        [Fact]
+        public async Task FetchConfig_InterceptedAnswer_FallsBack_WithoutCounting_AndReportsOnce()
+        {
+            // A 403 something on the network path answered (TLS inspection): same fallback as an
+            // auth failure, but the tracker does not count it — it sends ClientCertificateNotReceived.
+            var distress = new FakeDistressReporter();
+            var tracker = new AuthFailureTracker(maxFailures: 5, timeoutMinutes: 0,
+                AutopilotMonitor.DecisionCore.Engine.SystemClock.Instance, _logger, distress);
+            var svc = NewService(BackendBehaviour.AlwaysIntercepted403, tracker);
+
+            var result = await svc.FetchConfigAsync(retryOnTransientErrors: true);
+
+            Assert.NotNull(result);
+            Assert.Equal(RemoteConfigFetchOutcome.UsedDefaults, svc.LastFetchOutcome);
+            Assert.Equal(1, svc.LastFetchAttempts);
+            Assert.Equal(403, svc.LastFetchAuthStatusCode);
+            Assert.Equal(0, tracker.ConsecutiveFailures);
+            var call = Assert.Single(distress.Calls);
+            Assert.Equal(DistressErrorType.ClientCertificateNotReceived, call.ErrorType);
+            Assert.StartsWith("agent/config http 403 client_certificate_required", call.Message);
+        }
+
         [Fact]
         public async Task FetchConfig_RetryDisabled_OneAttemptOnly()
         {
@@ -144,7 +167,7 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Configuration
             Assert.Equal(4, svc.LastFetchAttempts);
         }
 
-        // ── Network-link gate (tenant aebdce78 audit 2026-08-24) ────────────
+        // ── Network-link gate (kiosk-tenant audit 2026-08-24) ────────────
 
         [Fact]
         public async Task FetchConfig_InitialFetch_WaitsForNetworkLink()
@@ -252,10 +275,11 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Configuration
             /// <summary>All 4 initial attempts fail; the NEXT call (the recovery fetch) succeeds.</summary>
             FailAllThenSucceed,
             AlwaysAuthFailure401,
+            AlwaysIntercepted403,
             AlwaysReturnNull,
         }
 
-        private FakeRemoteConfigService NewService(BackendBehaviour behaviour)
+        private FakeRemoteConfigService NewService(BackendBehaviour behaviour, AuthFailureTracker? tracker = null)
         {
             // Pass a throwaway BackendApiClient — the fake overrides CallBackendAsync
             // before it ever hits the wire.
@@ -269,7 +293,7 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Configuration
                 bootstrapToken: null,
                 agentVersion: "0.0.0",
                 logger: _logger);
-            return new FakeRemoteConfigService(apiClient, _logger, behaviour);
+            return new FakeRemoteConfigService(apiClient, _logger, behaviour, tracker);
         }
 
         private sealed class FakeRemoteConfigService : RemoteConfigService
@@ -277,8 +301,8 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Configuration
             private readonly BackendBehaviour _behaviour;
             private int _calls;
 
-            public FakeRemoteConfigService(BackendApiClient apiClient, AgentLogger logger, BackendBehaviour behaviour)
-                : base(apiClient, "00000000-0000-0000-0000-000000000001", logger)
+            public FakeRemoteConfigService(BackendApiClient apiClient, AgentLogger logger, BackendBehaviour behaviour, AuthFailureTracker? tracker)
+                : base(apiClient, "00000000-0000-0000-0000-000000000001", logger, authFailureTracker: tracker)
             {
                 _behaviour = behaviour;
             }
@@ -305,6 +329,9 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Configuration
                         throw new HttpRequestException("simulated network failure");
                     case BackendBehaviour.AlwaysAuthFailure401:
                         throw new BackendAuthException("Unauthorized", 401);
+                    case BackendBehaviour.AlwaysIntercepted403:
+                        throw new BackendAuthException("The device certificate did not reach the backend.", 403,
+                            classification: InterceptionFixtures.CertificateRequired());
                     case BackendBehaviour.AlwaysReturnNull:
                         return Task.FromResult<AgentConfigResponse>(null!);
                     default:

src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/Harness/FakeDistressReporter.cs

@@ -0,0 +1,42 @@
+using System;
+using System.Collections.Generic;
+using System.Threading.Tasks;
+using AutopilotMonitor.Agent.V2.Core.Monitoring.Transport;
+using AutopilotMonitor.Shared.Models;
+
+namespace AutopilotMonitor.Agent.V2.Core.Tests.Harness
+{
+    /// <summary>
+    /// Records every distress dispatch instead of sending it. The real anti-flood rules of
+    /// <see cref="DistressReporter"/> are bypassed on purpose: tests assert what the caller dispatched.
+    /// </summary>
+    internal sealed class FakeDistressReporter : DistressReporter
+    {
+        public sealed class Call
+        {
+            public DistressErrorType ErrorType { get; set; }
+            public string Message { get; set; } = default!;
+            public int? HttpStatusCode { get; set; }
+            public string? TpmVersion { get; set; }
+        }
+
+        public List<Call> Calls { get; } = new List<Call>();
+
+        /// <summary>
+        /// The send each dispatch returns; null = already completed. Lets a test keep a send in
+        /// flight to prove a caller waits for it (or stops waiting at its bound).
+        /// </summary>
+        public Func<Task>? Send { get; set; }
+
+        public FakeDistressReporter()
+            : base(baseUrl: "https://example.test", tenantId: "t", manufacturer: "m", model: "d", serialNumber: "s", agentVersion: "0.0.0", logger: null!)
+        {
+        }
+
+        public override Task TrySendAsync(DistressErrorType errorType, string message, int? httpStatusCode = null, string? tpmVersion = null)
+        {
+            Calls.Add(new Call { ErrorType = errorType, Message = message, HttpStatusCode = httpStatusCode, TpmVersion = tpmVersion });
+            return Send?.Invoke() ?? Task.CompletedTask;
+        }
+    }
+}

src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/Monitoring/Analyzers/SoftwareInventoryRegistryTextTests.cs

@@ -0,0 +1,31 @@
+#nullable enable
+using AutopilotMonitor.Agent.V2.Core.Monitoring.Telemetry.Analyzers;
+using Xunit;
+
+namespace AutopilotMonitor.Agent.V2.Core.Tests.Monitoring.Analyzers
+{
+    /// <summary>
+    /// Uninstall-key strings are read the way Windows shows them: up to the first NUL. Installers
+    /// that pad REG_SZ data with NULs otherwise leak them into the normalized vendor/name, which
+    /// the backend turns into a table key.
+    /// </summary>
+    public sealed class SoftwareInventoryRegistryTextTests
+    {
+        [Theory]
+        [InlineData("Contoso Widget", "Contoso Widget")]
+        [InlineData("Contoso Widget\0\0\0\0\0", "Contoso Widget")]
+        [InlineData("Contoso\0stale buffer", "Contoso")]
+        [InlineData("\0\0\0", "")]
+        [InlineData("", "")]
+        public void CutAtNul_KeepsTheTextBeforeTheFirstNul(string raw, string expected)
+        {
+            Assert.Equal(expected, SoftwareInventoryAnalyzer.CutAtNul(raw));
+        }
+
+        [Fact]
+        public void CutAtNul_Null_StaysNull()
+        {
+            Assert.Null(SoftwareInventoryAnalyzer.CutAtNul(null!));
+        }
+    }
+}

src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/Monitoring/EspAndHelloTrackerStarvedAppsTests.cs

@@ -174,7 +174,7 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Monitoring
         [Fact]
         public void Reevaluate_does_not_re_emit_starved_warnings()
         {
-            // sits-d Cloud-PC fix (2026-08-19): the synthesis re-check now runs on EVERY IME
+            // Cloud-PC fix (2026-08-19): the synthesis re-check now runs on EVERY IME
             // app-state transition. It must not turn the one-shot starvation warning into a
             // per-transition stream — the re-check path suppresses the emission entirely
             // (the per-appId dedupe would swallow the duplicates, but walking the probe
@@ -198,7 +198,7 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Monitoring
         [Fact]
         public void Starved_apps_that_settle_later_still_complete_the_enrollment()
         {
-            // The real sits-d shape: the ESP exits with apps still starved (warning emitted),
+            // The real field shape: the ESP exits with apps still starved (warning emitted),
             // the apps then all settle, and the re-check must open the gate — the warning
             // already on the timeline is history, not a permanent veto.
             using var f = new Fixture();

src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/Monitoring/EspAndHelloTrackerUserAppsSettledSynthesisTests.cs

@@ -172,8 +172,8 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Monitoring
         }
 
         // ------------------------------------------------------------------------------
-        // sits-d Cloud-PC fix (2026-08-19) — the synthesis is no longer edge-only.
-        // Sessions 8110e262 / a89aac2d: the ESP page exited while 138 required user-ESP apps
+        // Cloud-PC fix (2026-08-19) — the synthesis is no longer edge-only.
+        // Two field sessions: the ESP page exited while 138 required user-ESP apps
         // were still in flight, they all reached a terminal state (0 failed) minutes later,
         // and because nothing re-checked, the enrollment never completed.
         // ------------------------------------------------------------------------------

src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/Monitoring/NetworkFingerprintTests.cs

@@ -0,0 +1,85 @@
+using System.Linq;
+using AutopilotMonitor.Agent.V2.Core.Monitoring;
+using AutopilotMonitor.Agent.V2.Core.Monitoring.Telemetry.Periodic;
+using Xunit;
+
+namespace AutopilotMonitor.Agent.V2.Core.Tests.Monitoring
+{
+    /// <summary>
+    /// The local network fingerprint a failed registration is stored with: SHA-256 over the
+    /// connection type and gateway <see cref="NetworkChangeDetector"/> captures, cut to 16 hex
+    /// characters; unknown (null) whenever either input is missing. Gateways are documentation
+    /// addresses (RFC 5737).
+    /// </summary>
+    public sealed class NetworkFingerprintTests
+    {
+        private static NetworkChangeDetector.NetworkStateSnapshot State(string? connectionType, string? gateway, bool hasNetwork = true) =>
+            new NetworkChangeDetector.NetworkStateSnapshot
+            {
+                HasNetwork = hasNetwork,
+                ConnectionType = connectionType,
+                Gateway = gateway,
+            };
+
+        [Fact]
+        public void Same_connection_type_and_gateway_give_the_same_short_hex_fingerprint()
+        {
+            var first = NetworkFingerprint.Of(State("WiFi", "192.0.2.1"));
+            var second = NetworkFingerprint.Of(State("WiFi", "192.0.2.1"));
+
+            Assert.Equal(first, second);
+            Assert.Equal(NetworkFingerprint.Length, first!.Length);
+            Assert.True(first.All(c => (c >= '0' && c <= '9') || (c >= 'a' && c <= 'f')), first);
+        }
+
+        [Theory]
+        [InlineData("WiFi", "198.51.100.1")] // another gateway
+        [InlineData("Ethernet", "192.0.2.1")] // another connection type
+        public void Another_gateway_or_connection_type_is_another_network(string connectionType, string gateway)
+        {
+            Assert.NotEqual(
+                NetworkFingerprint.Of(State("WiFi", "192.0.2.1")),
+                NetworkFingerprint.Of(State(connectionType, gateway)));
+        }
+
+        [Fact]
+        public void Unknown_network_has_no_fingerprint()
+        {
+            Assert.Null(NetworkFingerprint.Of(null));
+            Assert.Null(NetworkFingerprint.Of(State("WiFi", "192.0.2.1", hasNetwork: false)));
+            Assert.Null(NetworkFingerprint.Of(State("WiFi", null)));
+            Assert.Null(NetworkFingerprint.Of(State(null, "192.0.2.1")));
+        }
+
+        [Theory]
+        [InlineData("aaaa", "aaaa", false)]
+        [InlineData("aaaa", "bbbb", true)]
+        public void Differs_compares_known_fingerprints(string before, string after, bool expected)
+        {
+            Assert.Equal(expected, NetworkFingerprint.Differs(before, after));
+        }
+
+        [Theory]
+        [InlineData(null, "aaaa")]
+        [InlineData("aaaa", null)]
+        [InlineData(null, null)]
+        public void Differs_is_unknown_when_either_side_is(string? before, string? after)
+        {
+            Assert.Null(NetworkFingerprint.Differs(before, after));
+        }
+
+        [Fact]
+        public void TryCapture_reads_this_machine_without_throwing()
+        {
+            // Machine-dependent: either no active interface (null) or a well-formed fingerprint
+            // that matches the detector's own capture.
+            var captured = NetworkFingerprint.TryCapture();
+
+            if (captured != null)
+            {
+                Assert.Equal(NetworkFingerprint.Length, captured.Length);
+                Assert.Equal(NetworkFingerprint.Of(NetworkChangeDetector.CaptureInterfaceState()), captured);
+            }
+        }
+    }
+}

src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/Monitoring/Runtime/BootstrapProvenanceReaderTests.cs

@@ -0,0 +1,199 @@
+using System;
+using System.Collections.Generic;
+using System.IO;
+using System.Linq;
+using System.Text;
+using AutopilotMonitor.Agent.V2.Core.Monitoring.Runtime;
+using Xunit;
+
+namespace AutopilotMonitor.Agent.V2.Core.Tests.Monitoring.Runtime
+{
+    /// <summary>
+    /// The bootstrap provenance in <c>agent_started</c>. Fixture lines are the scripts' real
+    /// formats: installer <c>[ts] Bootstrap script version: vX.Y</c>
+    /// (Install-AutopilotMonitor.ps1), loader <c>[ts] === Autopilot Monitor bootstrap loader vX.Y
+    /// started (user=..., pid=...) ===</c> (Start-AutopilotMonitor.ps1), the MSI's downloader before
+    /// the loader <c>[ts] MSI bootstrap runner started (user=..., pid=...).</c>.
+    /// </summary>
+    public sealed class BootstrapProvenanceReaderTests : IDisposable
+    {
+        private readonly string _dir = Path.Combine(Path.GetTempPath(), "apm-provenance-" + Guid.NewGuid().ToString("N"));
+
+        public BootstrapProvenanceReaderTests() => Directory.CreateDirectory(_dir);
+
+        public void Dispose()
+        {
+            try { Directory.Delete(_dir, recursive: true); } catch { /* best effort */ }
+        }
+
+        private void Write(string name, params string[] lines) =>
+            File.WriteAllText(Path.Combine(_dir, name), string.Join("\r\n", lines) + "\r\n", Encoding.ASCII);
+
+        private static string InstallerRun(string version) => $"[2026-10-13 09:00:00] Bootstrap script version: v{version}";
+
+        private static string LoaderRun(string version) =>
+            $"[2026-10-13 08:59:58] === Autopilot Monitor bootstrap loader v{version} started (user=SECRETUSER, pid=4242) ===";
+
+        [Fact]
+        public void Loader_path_reports_installer_and_loader_versions()
+        {
+            Write(BootstrapProvenanceReader.LoaderLogName, LoaderRun("1.2"), "[2026-10-13 08:59:59] Installer signature verified");
+            Write(BootstrapProvenanceReader.InstallerLogName, "[2026-10-13 09:00:00] ===== Autopilot Monitor Bootstrap Started =====", InstallerRun("2.5"));
+
+            var p = BootstrapProvenanceReader.Read(_dir);
+
+            Assert.True(p.ScriptLogPresent);
+            Assert.Equal("2.5", p.ScriptVersion);
+            Assert.Equal("loader", p.LoaderLog);
+            Assert.Equal("1.2", p.LoaderVersion);
+            Assert.False(p.OobeLogPresent);
+        }
+
+        [Fact]
+        public void The_last_run_wins_when_the_log_grew_over_several_runs()
+        {
+            Write(BootstrapProvenanceReader.InstallerLogName, InstallerRun("2.3"), "x", InstallerRun("2.4"), "y", InstallerRun("2.5"), "z");
+
+            Assert.Equal("2.5", BootstrapProvenanceReader.Read(_dir).ScriptVersion);
+        }
+
+        [Fact]
+        public void A_directly_assigned_installer_has_no_loader_log()
+        {
+            Write(BootstrapProvenanceReader.InstallerLogName, InstallerRun("2.4"));
+
+            var p = BootstrapProvenanceReader.Read(_dir);
+
+            Assert.Equal("2.4", p.ScriptVersion);
+            Assert.Null(p.LoaderLog);
+            Assert.Null(p.LoaderVersion);
+        }
+
+        [Fact]
+        public void An_installer_from_before_the_version_line_is_present_but_without_version()
+        {
+            Write(BootstrapProvenanceReader.InstallerLogName, "[2026-03-01 09:00:00] ===== Autopilot Monitor Bootstrap Started =====");
+
+            var p = BootstrapProvenanceReader.Read(_dir);
+
+            Assert.True(p.ScriptLogPresent);
+            Assert.Null(p.ScriptVersion);
+        }
+
+        [Fact]
+        public void An_installer_from_before_the_log_rename_is_read_under_its_old_name()
+        {
+            Write(BootstrapProvenanceReader.LegacyInstallerLogName, InstallerRun("2.3"));
+
+            var p = BootstrapProvenanceReader.Read(_dir);
+
+            Assert.True(p.ScriptLogPresent);
+            Assert.Equal("2.3", p.ScriptVersion);
+        }
+
+        [Fact]
+        public void With_both_installer_logs_the_most_recently_written_one_counts()
+        {
+            Write(BootstrapProvenanceReader.InstallerLogName, InstallerRun("2.6"));
+            Write(BootstrapProvenanceReader.LegacyInstallerLogName, InstallerRun("2.3"));
+            File.SetLastWriteTimeUtc(Path.Combine(_dir, BootstrapProvenanceReader.InstallerLogName), DateTime.UtcNow.AddDays(-3));
+            File.SetLastWriteTimeUtc(Path.Combine(_dir, BootstrapProvenanceReader.LegacyInstallerLogName), DateTime.UtcNow.AddMinutes(-1));
+
+            Assert.Equal("2.3", BootstrapProvenanceReader.Read(_dir).ScriptVersion);
+        }
+
+        [Fact]
+        public void The_msi_downloader_from_before_the_loader_counts_as_msi_without_version()
+        {
+            Write(BootstrapProvenanceReader.LegacyMsiLoaderLogName, "[2026-08-01 09:00:00] MSI bootstrap runner started (user=SECRETUSER, pid=4242).");
+
+            var p = BootstrapProvenanceReader.Read(_dir);
+
+            Assert.Equal("msi", p.LoaderLog);
+            Assert.Null(p.LoaderVersion);
+        }
+
+        [Fact]
+        public void The_msi_loader_log_is_told_apart_from_the_intune_loader()
+        {
+            Write(BootstrapProvenanceReader.MsiLoaderLogName, LoaderRun("1.1"));
+
+            var p = BootstrapProvenanceReader.Read(_dir);
+
+            Assert.Equal("msi", p.LoaderLog);
+            Assert.Equal("1.1", p.LoaderVersion);
+        }
+
+        [Fact]
+        public void With_both_loader_logs_the_most_recently_written_one_counts()
+        {
+            Write(BootstrapProvenanceReader.LoaderLogName, LoaderRun("1.0"));
+            Write(BootstrapProvenanceReader.MsiLoaderLogName, LoaderRun("1.2"));
+            File.SetLastWriteTimeUtc(Path.Combine(_dir, BootstrapProvenanceReader.LoaderLogName), DateTime.UtcNow.AddDays(-3));
+            File.SetLastWriteTimeUtc(Path.Combine(_dir, BootstrapProvenanceReader.MsiLoaderLogName), DateTime.UtcNow.AddMinutes(-1));
+
+            var p = BootstrapProvenanceReader.Read(_dir);
+
+            Assert.Equal("msi", p.LoaderLog);
+            Assert.Equal("1.2", p.LoaderVersion);
+        }
+
+        [Fact]
+        public void The_oobe_go_script_is_recognised_by_its_log()
+        {
+            Write(BootstrapProvenanceReader.OobeLogName, "[2026-10-13 09:00:00] ===== Autopilot Monitor OOBE Bootstrap =====");
+
+            Assert.True(BootstrapProvenanceReader.Read(_dir).OobeLogPresent);
+        }
+
+        [Fact]
+        public void Only_the_tail_of_a_large_log_is_read()
+        {
+            var filler = Enumerable.Repeat(new string('.', 200), 600).ToArray();   // ~120 KB
+            var lines = new List<string> { InstallerRun("2.1") };
+            lines.AddRange(filler);
+            lines.Add(InstallerRun("2.5"));
+            Write(BootstrapProvenanceReader.InstallerLogName, lines.ToArray());
+
+            Assert.Equal("2.5", BootstrapProvenanceReader.Read(_dir).ScriptVersion);
+        }
+
+        [Fact]
+        public void A_missing_directory_reads_as_unknown_and_never_throws()
+        {
+            var p = BootstrapProvenanceReader.Read(Path.Combine(_dir, "does-not-exist"));
+
+            Assert.False(p.ScriptLogPresent);
+            Assert.Null(p.ScriptVersion);
+            Assert.Null(p.LoaderLog);
+            Assert.False(p.OobeLogPresent);
+        }
+
+        [Fact]
+        public void The_event_keys_carry_versions_only_never_the_account_name()
+        {
+            Write(BootstrapProvenanceReader.LoaderLogName, LoaderRun("1.2"));
+            Write(BootstrapProvenanceReader.InstallerLogName, InstallerRun("2.5"));
+            var data = new Dictionary<string, object>();
+
+            BootstrapProvenanceReader.Read(_dir).AddTo(data);
+
+            Assert.Equal(true, data["bootstrapScriptLogPresent"]);
+            Assert.Equal("2.5", data["bootstrapScriptVersion"]);
+            Assert.Equal("loader", data["bootstrapLoaderLog"]);
+            Assert.Equal("1.2", data["bootstrapLoaderVersion"]);
+            Assert.Equal(false, data["oobeBootstrapLogPresent"]);
+            Assert.DoesNotContain(data.Values, v => v is string s && s.Contains("SECRETUSER"));
+        }
+
+        [Fact]
+        public void Keys_that_were_not_found_are_left_out()
+        {
+            var data = new Dictionary<string, object>();
+
+            BootstrapProvenanceReader.Read(_dir).AddTo(data);
+
+            Assert.Equal(new[] { "bootstrapScriptLogPresent", "oobeBootstrapLogPresent" }, data.Keys.OrderBy(k => k).ToArray());
+        }
+    }
+}

src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/Monitoring/Runtime/SelfUpdaterPayloadCatalogTests.cs

@@ -0,0 +1,101 @@
+using System;
+using System.Collections.Generic;
+using System.IO;
+using AutopilotMonitor.Agent.V2.Core.Monitoring.Runtime;
+using AutopilotMonitor.Agent.V2.Core.Security;
+using Xunit;
+
+namespace AutopilotMonitor.Agent.V2.Core.Tests.Monitoring.Runtime
+{
+    /// <summary>
+    /// Step 5b of the self-update: the staged payload must match its signed catalog before a
+    /// single file is swapped in. The verifier is replaced through the static hook; each test
+    /// restores it, as in <see cref="SelfUpdaterGracefulShutdownTests"/>.
+    /// </summary>
+    public sealed class SelfUpdaterPayloadCatalogTests : IDisposable
+    {
+        private const string StagingDir = @"C:\ProgramData\AutopilotMonitor\Agent-Update";
+
+        private readonly Func<string, PayloadCatalogVerdict> _savedVerifier;
+        private readonly List<string> _log = new List<string>();
+
+        public SelfUpdaterPayloadCatalogTests()
+        {
+            _savedVerifier = SelfUpdater.VerifyPayloadCatalog;
+        }
+
+        public void Dispose() => SelfUpdater.VerifyPayloadCatalog = _savedVerifier;
+
+        [Fact]
+        public void Production_default_is_the_real_verifier()
+        {
+            Assert.Equal(typeof(PayloadCatalogVerifier), _savedVerifier.Method.DeclaringType);
+            Assert.Equal(nameof(PayloadCatalogVerifier.Verify), _savedVerifier.Method.Name);
+        }
+
+        [Fact]
+        public void Accepted_payload_lets_the_update_continue()
+        {
+            string? checkedDir = null;
+            SelfUpdater.VerifyPayloadCatalog = dir => { checkedDir = dir; return PayloadCatalogVerdict.Accept(); };
+
+            var ok = SelfUpdater.VerifyStagedPayload(StagingDir, _log.Add, out var reason);
+
+            Assert.True(ok);
+            Assert.Null(reason);
+            Assert.Equal(StagingDir, checkedDir);
+            Assert.Contains(_log, line => line.Contains("payload catalog verified"));
+        }
+
+        [Fact]
+        public void Rejected_payload_stops_the_update_and_hands_over_the_reason()
+        {
+            SelfUpdater.VerifyPayloadCatalog = _ =>
+                PayloadCatalogVerdict.Reject("payload differs from catalog (missing: none; extra: evil.dll; changed: none)");
+
+            var ok = SelfUpdater.VerifyStagedPayload(StagingDir, _log.Add, out var reason);
+
+            Assert.False(ok);
+            Assert.Contains("extra: evil.dll", reason);
+            Assert.Contains(_log, line => line.Contains("REJECTED"));
+        }
+
+        [Fact]
+        public void A_throwing_verifier_is_a_rejection_not_a_pass()
+        {
+            SelfUpdater.VerifyPayloadCatalog = _ => throw new InvalidOperationException("runspace broke");
+
+            var ok = SelfUpdater.VerifyStagedPayload(StagingDir, _log.Add, out var reason);
+
+            Assert.False(ok);
+            Assert.Contains("runspace broke", reason);
+        }
+
+        [Fact]
+        public void The_catalog_check_runs_before_any_file_is_swapped()
+        {
+            // Source-level pin: CheckAndApplyUpdateAsync has no seam for the download, so the
+            // ordering that makes the check meaningful is pinned on the source itself.
+            var source = File.ReadAllText(Path.Combine(FindRepoRoot(),
+                "src", "Agent", "AutopilotMonitor.Agent.V2.Core", "Monitoring", "Runtime", "SelfUpdater.cs"));
+
+            var check = source.IndexOf("if (!VerifyStagedPayload(stagingDir, log, out var catalogRejection))", StringComparison.Ordinal);
+            var swap = source.IndexOf("var swapOk = SwapFiles(agentDir, stagingDir, log);", StringComparison.Ordinal);
+
+            Assert.True(check > 0, "step 5b not found in SelfUpdater.cs");
+            Assert.True(swap > check, "SwapFiles must come after the catalog check");
+            Assert.Contains("WriteSkipMarker(\"catalog_rejected\"", source);
+        }
+
+        private static string FindRepoRoot()
+        {
+            var dir = new DirectoryInfo(AppContext.BaseDirectory);
+            while (dir != null && !File.Exists(Path.Combine(dir.FullName, "AutopilotMonitor.sln")))
+            {
+                dir = dir.Parent;
+            }
+            Assert.NotNull(dir);
+            return dir.FullName;
+        }
+    }
+}

src/Agent/AutopilotMonitor.Agent.V2.Core.Tests/Monitoring/ShellCoreTrackerBreadcrumbTests.cs

@@ -11,6 +11,7 @@ using AutopilotMonitor.Agent.V2.Core.Tests.Orchestration;
 using AutopilotMonitor.DecisionCore.Engine;
 using AutopilotMonitor.DecisionCore.Signals;
 using AutopilotMonitor.Shared;
+using AutopilotMonitor.Shared.Models;
 using Newtonsoft.Json;
 using Xunit;
 
@@ -173,16 +174,16 @@ namespace AutopilotMonitor.Agent.V2.Core.Tests.Monitoring
         public void UpdatePageRecords_ArePageStartsStopsAndNames_OfTheUpdatePageOnly()
         {
             var start = ShellCoreTracker.ParseUpdatePageRecord(ShellCoreTracker.EventId_ShellCore_WebAppStarted, NdupStarted, lastPage: null);
-            Assert.Equal(OobeUpdatePageRecord.PageStarted, start!.CxhEvent);
+            Assert.Equal(OobeUpdateVocabulary.PageStarted, start!.CxhEvent);
             Assert.Equal("OobeNDUP", start.Page);
 
             var stop = ShellCoreTracker.ParseUpdatePageRecord(ShellCoreTracker.EventId_ShellCore_WebAppStopped, PageStopped, lastPage: "RebootNDUP");
-            Assert.Equal(OobeUpdatePageRecord.PageStopped, stop!.CxhEvent);
+            Assert.Equal(OobeUpdateVocabulary.PageStopped, stop!.CxhEvent);
             Assert.Equal("RebootNDUP", stop.Page);
             Assert.Equal("success", stop.Result);
 
             var name = ShellCoreTracker.ParseUpdatePageRecord(ShellCoreTracker.EventId_ShellCore_WebAppEventName, ScanStartedName, lastPage: null);
-            Assert.Equal(OobeUpdatePageRecord.EventName, name!.CxhEvent);
+            Assert.Equal(OobeUpdateVocabulary.EventName, name!.CxhEvent);
             Assert.Equal("ExpeditedUpdate_startWUScanStarted", name.Name);
             Assert.Null(name.Value);