Microsoft switched off the Office 365 connectors in Teams in May 2026. Provider 1,
its MessageCard renderer, the channel editor option and the flat TeamsWebhookUrl /
TeamsNotifyOn* fields with their fallback branch are gone; the value stays retired.
NotificationChannel.IsSupportedProviderType is now the one rule for channel parsing,
channel validation and the flat WebhookProviderType, so a retired or unknown provider
can no longer be stored and come back as a synthesized channel. The channel editor
looks up the placeholder by value instead of by list index. The header-parsing tests
move to NotificationChannel.GetCustomHeaders; manifests and wire types regenerated.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The time attribution shows the OOBE quality update only once a session ended. While it
runs, the web now derives it from the events it already holds:
- lib/osUpdateLive.ts ports the calculator's update interval for the current window and
labels it from positive evidence that the update finished: running (restarting),
waiting for sign-in from 2 min after its end, or failed/skipped without a restart.
A package waiting for its restart keeps the update running whatever the page reported.
- Session page: a line under the phase timeline, shown after the first complete walk of
the event pages, with a note when the device has been silent for 15 minutes.
- Progress Portal: the headline says that Windows is installing updates and asks the user
to sign in after the update's restart.
- OobeUpdateVocabulary (Shared) is the one vocabulary of the agent, the calculator and the
web (manifest section oobeUpdate); same strings on the wire, no behaviour change.
- tests/fixtures/os-update-live/cases.json pins the calculator and the port
(TimeAttributionOsUpdateLiveParityTests, lib/__tests__/osUpdateLive.test.ts).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- An X click counts a dismissal (DismissCount on the person's row); the second ask waits for the
configured pause (FeedbackCooldownDays), the second dismissal ends the asking. A rating ends it at
once. Ignoring the bubble records nothing, as before.
- A rated row is final: a later dismissal or rating (another tab) changes nothing and sends no second
ping. Answers are written conditionally on what was read (ETag), so two tabs cannot overwrite a rating.
- Rows from before the count read as one dismissal when dismissed; the person keeps UPN and name.
- Rules in FeedbackPromptRules; the admin Ratings tab shows "dismissed twice".
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- POST feedback/general (MemberRead): text up to 4096 characters, optional reply address, at most
10 per user in 24 hours; one Feedback row per message (PK General), shown in a new admin tab.
- FeedbackReceived ops event without text or sender, plus a bell entry. The help menu gets one
"Send feedback" entry; the dialog links to Submit Logs for Admins and Operators.
- Report uploads: logs are zipped before sending and both forms share one budget derived from the
20 MiB body cap (Constants.SubmissionLimits through the shared manifest). The session exports go
along when they fit next to the attachments.
- Report Session opens to Operators. Comment and contact address are validated; a report whose row
cannot be written fails and its uploaded blobs are removed.
- feedback/status no longer creates a tenant configuration row for an unknown tenant.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The first run has no Shell-Core replay (the restart replay re-raises the Hello-wizard start), so an
agent installed late during the ESP missed the update page's start and first half (6a7d241d).
- BackfillUpdatePageTelemetry: on the first run the last 60 min of Shell-Core become oobe_update_page
rows and nothing else — no Hello replay, no ESP exit or failure, no agent_trace, no registry read.
- Live and backfilled names have budgets of 150 each; the first name over a budget sends a
names_capped marker, so a missing name is never read as "did not happen".
- A backfill batch keeps its own last page and moves the live watcher's only with a newer RecordId:
a restart replay's older page start no longer takes over the next live stop, and a first run's
live stop finds the page that started before the agent.
Agent suite 3,436 green; the agent stays on net48.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Field measurement 3: a first-run agent that started after the update page (6a7d241d) left the whole
update in Apps (ESP), failed and skipped updates were reported by nothing, and the lane listed a
failed update's staged servicing stack as an installed KB.
- Time attribution v5: update names before the first observed OobeNDUP start begin the update at the
earlier of the first name and the update's first servicing step (after the window's ESP exit when
one was seen before the name, else the first step of a package the update also serviced later).
The fallback's ESP exit is per window, so a WhiteGlove part-1 exit no longer begins an update in
part 2. OsUpdateSpan carries Outcome (installed/failed/skipped/unknown, the latest evidence wins)
and NotInstalledKbs; Kbs lists only packages that reached Installed. The 2-h sweep recomputes the
last 30 days.
- ANALYZE-DEV-013 (update failed), DEV-014 (not offered after an earlier failure), DEV-015 (update
check ended without a result): warning, never KO; a failure only from explicit names, every
absence condition gated on the agent's names_capped marker; never two at once.
- ShellCoreTracker rows never set the session start (D-315 list), ahead of the agent's first-run
update page backfill.
- The web lane names the outcome and the KBs not installed; MCP get_time_attribution mentions both.
Real-event replay: 6a7d241d os_update 2358 s installed (was 55 min in esp_apps), e7551b82 failed with
KB5124007 not installed. Rule dry runs and a sweep over all 97 update-page sessions since 2026-10-02
hit exactly the five field sessions. Functions 6,918, web 1,701, MCP 997 green.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Tailwind 3's divide-<color> rule (`> :not([hidden]) ~ :not([hidden])`,
specificity 0,3,0, plus the dark-mode !important overrides in globals.css)
recolours every child but the first on all sides, so a row's own border
accent only ever showed on the first row.
- Notification dropdown: the purple GA and blue tenant bars are inset
shadows now, so every such row keeps its bar.
- Geographic performance: the red outlier bar moves from the row to its
first cell; it was grey on every outlier but the first table row.
- Tenant config report: the heavier separator between notification
channels and between self-hosted AI clients moves to the cell
(GROUP_SEPARATOR); the divider had shrunk it to a plain 1px line.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
A concurrent IME writer overwrites lines behind the bookmark, and the overwrite check recovers them a
pass later. Recovered lines were never fresh, so they resolved in the agent's own cached zone: in a lab
the zone was stale (+60, IME wrote +120), recovered result lines landed an hour ahead of the clock and
the platform-script durations were omitted ("n/a" on Install Autopilot Monitor, session d29e27e3, also
on 2.0.1474/2.0.1467). Bytes that read differently at a position were written after that position was
last seen, and IME stamps a line when it writes it, so the last sighting bounds the new line's age.
- Ledger entries carry SeenUtc: set when read or seeded, renewed whenever a verification finds them
unchanged.
- A changed entry below the old bookmark is fresh when the pass is, the ledger held the entry at the
rewind point, and the replaced bytes were seen 0-30 s ago; otherwise the fallback as before.
Re-processed entries are never calibration anchors.
Field replay and longer-running script red before; six mutation proofs; agent suite 3,426 green.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
IME re-runs a failed platform script on later check-ins, and the agent's first pass reads all of
AgentExecutor.log before the IME log, so several runs of one policy shared one slot: the first IME
result took the LAST run's exit code and output (session eff1b413: the bootstrap script "Failed
(exit 0)" with the stdout of its successful retry), every later result went out without an end
block. The exit-code fallback said Success for exit 0 with stderr and Failed for exit 1 without,
which IME never does.
- A later timestamped executor start parks the previous run of its policy (persisted, max 64); an
IME result pairs with the run of its policy and user that started last before the result line;
earlier parked runs of the same user resolve then, so emits keep run order.
- Invocation markers carry the run id; the emitted marker only rises; no dedup path deletes a live
run; a parked emit never takes the later run's D-316 register entry.
- Fallback: Success exactly when the stderr is empty, the exit code only when no stderr line was
read; "write output done … error =" replaces the "error from script" snapshot, empty stdout too.
Backfill replays fail on the old code; nine mutation proofs; agent suite 3,416 green.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Web: the script panel applies the correction to the run its runId names: the saved output replaces the emitted one, a foreign end block's exit code shows as n/a with an explanation, the state follows the corrected data instead of the emitted event type, and the output carries a short "from IME's saved result" note. The timeline marks the original row as corrected later, and the correction event falls under the script-output setting like the other script events. The session page passes the event to the panel; the gather-rule event list knows the type.
MCP: get_session_summary no longer counts a script_failed as an error, ranks it as one or attaches the exit code's catalog text when a foreign correction with IME result Success names its run.
Backend: the correction never heals a stalled session and never counts as a script run.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
After reporting a script batch, IME saves every run under HKLM\SOFTWARE\Microsoft\IntuneManagementExtension\Policies\<user>\<policy>; ResultDetails carries the stdout IME read from that policy's own result file (the stderr when it failed with error code 99). The agent now checks each emitted platform run against it: a match confirms the output and the exit code of the same end block, a contradiction emits script_output_reconciliation (Warning, outcome "foreign": the saved output replaces the emitted one, the exit code is unverified), and output the log delivered incomplete is completed (Info, outcome "repaired").
A save is paired with the latest run of its user and policy only after the tracker completed a pass that started two seconds after the save was seen, so a stalled poll loop cannot pair it with an older run. Platform completions carry a runId the correction names. Two restart-safe counters (verified, corrected) ride agent_metrics_snapshot and ime_pattern_hits. The new event type is registered in the shared constants, the generated manifests and the MCP event catalog.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Two AgentExecutor processes that open the log at the same end write their start blocks over each other. The end block of the one whose start block lost then sat behind the other executor's start block, and position-based ownership handed a Win32 detection script's stdout and exit code to the platform script.
An AgentExecutor.log entry that is not exactly one CMTrace record (a fragment, a merged line or an empty line) is where one writer's bytes end and another's begin: it now closes the platform invocation in front of it and is never matched. Invocation markers live as long as the bytes of their entry: a rewind sets them aside and gives an entry that reads unchanged its markers back.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Since agent 2.0.1473 the Windows Update and servicing watchers backfill rows with their event-log
time up to 60 min before the agent started, and the session anchor only skipped the system
timeline watcher, so session starts and durations grew by 22 min at the median.
- Shared: Constants.EventSources.SessionAnchorIneligible and IsSessionAnchorEligible are the one
rule; the agent writes the same constants as Source.
- Backend: the ingest batch (ClassifyEvents, now internal static) and the Events-table probe use it.
- Web: the session duration reads the list from shared-manifests.json.
- Recorded sessions keep their stored start; the rule applies to new ingests.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
GitHub masks only secrets.* by itself. The bearer token Get-StorageAuthHeaders.ps1
mints at run time was never registered as a mask, unlike the Key Vault signing token,
which build-agent, build-msi and publish-scripts mask before exporting it. No caller
prints it today -- every call site captures the hashtable and passes it only as
-Headers -- but this repository's run logs are public, and a single debugging line
would have published it: an uncaptured call of the helper, or Get-Error /
Format-List on a failed request, whose TargetObject is the HttpRequestMessage
including the Authorization header (reproduced locally with a dummy token).
The helper now emits ::add-mask:: for every token it hands out, which covers all six
workflows in one place. Write-Host, never Write-Output: the output stream is the
script's return value. Verified with a mocked az that the result stays one hashtable
and that a nested caller in the Request-BuildNumber.ps1 pattern still returns only
its build number. Only under GITHUB_ACTIONS -- locally the line would print the token
instead of hiding it.
Prompted by an external AI review of the public scripts.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- The Members card was the only card in the edit modal with a separate
icon column: its title, add form and member list sat 32px in from the
card edge, which on a phone also took 32px from the email field and the
member rows.
- It now uses the layout of its sibling cards: the title on top, the
content at full card width.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HNLaZbFQTdhQJU9jh37Hn6
Since e537a4b the backend writes a new versioned file per publish, so its
immutable header holds again. The browser revalidates only the manifest
(no-cache) and serves the file it names from its cache (default mode),
instead of revalidating both, the workaround in 762fef4 for the per-day name.
The build log names the baked file ("Platform stats baked into the landing
page from platform-stats.<date>T<time>Z.json (as of ...)"), so every web
deploy shows which publish it carries.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N4ddfosgdeCwXo7NvB9xd2
The versioned stats file is served public, max-age=31536000, immutable, and
the manifest platform-stats.json names the current one. Its per-day name
(19e9713a) was unique while maintenance ran once a night; since 82919d47 it
runs every 2 hours and rewrote that file up to 12 times a day, so caches kept
the first copy of the day. The name now carries the publish instant to the
second in invariant culture (platform-stats.2026-10-03T060004Z.json).
Headers and access tier travel with the upload instead of two follow-up
calls, which left the blob briefly without Cache-Control and content type,
and kept it that way (with the manifest still on the old file) when one of
them failed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N4ddfosgdeCwXo7NvB9xd2
Crawlers and LLM fetchers run no JS and saw only the skeleton of the stats
band. app/page.tsx now awaits one read of the public stats blob during
next build (lib/platformStats.ts: manifest, then the versioned file it
names) and renders the numbers into the export. StatsBand starts from that
snapshot and replaces it with the live numbers; a failed live read keeps
the baked numbers instead of hiding the band.
The read is awaited at page level, not under Suspense: React moves a
resolved boundary of a page this size into a hidden div at the end of the
body. It uses no-store (route force-static), so Next's fetch cache cannot
hand a later build the previous numbers. Fail-soft: without a snapshot the
page ships the skeleton and loads live as before, and deploy-web.yml warns.
The browser now revalidates both files (no-cache instead of force-cache):
the backend rewrites the per-day file every 2 hours under immutable, so a
copy cached earlier that day would have made the numbers drop after load.
The page body moved to components/landing/LandingPage.tsx, which the
marketing guards render synchronously; StatsBand left the .json() ratchet
baseline; the sitemap lastmod of / includes the body.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N4ddfosgdeCwXo7NvB9xd2
TimeAttributionCalculator v4 cuts two segments out of the phase partition:
os_update from the OobeNDUP page start whose visit holds update evidence
(a servicing step or a name only an update in progress writes) to the last
update evidence before the user is back; awaiting_sign_in from there to the
first evidence of the user, only after a restart. Both replace the phase
span they overlap; without evidence the partition is v3's. OsUpdates names
the KBs and restarts. Reboot gaps no longer start or end on the clock-sync
row logged at boot. The sweep recomputes the last 30 days.
Aggregates carry eight segments; the session lane and Fleet Health show
"Windows Update" (cyan) and "Waiting for sign-in" (sky) with the KBs and
restarts in legend and tooltip; MCP get_time_attribution description;
manifest and wire types regenerated.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
ShellCoreTracker reads CXID, result, name and value from the event properties
(62404..62407 %1/%2), live and in the backfill. The formatted message is
localized: German devices lost result and value, French devices reported no
update page at all. The text patterns stay as the fallback and accept
guillemets. The Hello-wizard start reads the CXID property too.
OobeUpdateTelemetry: page start and stop sit outside the run budget (at most
3 per page), names share a budget of 150 (was 50; the cloud page logs ~45
init names in its first 40 s). ReportPage returns whether it reported.
A RecordId watermark (shellcore-page-watermark.json) claims every update-page
record and persists reported ones, so the restart backfill does not send the
minutes before the restart again. EspAndHelloTracker passes the state
directory.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The form no longer lays out by viewport breakpoint: a wide window can host a
narrow column next to the settings sidebar, where the input shrank to a few
characters. The input now grows from an 18rem basis and wraps below the type
switch when the row has no room; role and Add wrap together.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- An open member row keeps role and bootstrap permission as a draft:
Save writes it, Discard or closing the row drops it, so a stray select
change no longer alters anyone's access.
- AddMemberForm takes `stacked` for narrow hosts; the GA tenant editor
modal uses it, so the email field spans the full width there.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- EntraAppRoleResolver maps the Viewer app role next to Admin and Operator
(the highest role wins); a claim Viewer gets no bootstrap-token rights.
- The tenant settings page and the GA tenant editor share components/members:
one line per member, a click opens who added it and the actions, Remove
asks in the row, role chips with counts, a search, 25 rows per page. A
vanished chip or a shrunken list moves the selection, so the list never
shows an empty page and never snaps back.
- The add form sits above the list, About Roles is collapsed, and a tenant
with Entra app roles enabled sees that those users are not listed.
- The EntraAppRolesEnabled doc names Viewer; manifest and generated wire
types regenerated.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
CodeQL js/incomplete-hostname-regexp (alert #96) traced DOCS_URL into the
RegExp built in the "Learn more" assertion. The dots were escaped with
[.], which CodeQL does not recognize as a sanitizer. A substring match
on the exact markup needs no escaping and keeps the same checks: a space
after the sentence, the exact href with target/rel, and "Learn more" as
the link text.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
GitBook publishes a page under the slug of its SUMMARY.md group, not its
folder: troubleshooting/ is /troubleshooting-and-support/, trust/ is
/trust-and-security/. A folder URL works only for pages that predate the
group rename (307); for a newer page it is a 404.
- Web: the Progress Portal access hint linked the new troubleshooting
article by folder path (404 live). It now uses DOCS_PATHS.progressPortalOnly.
- MCP: docUrl built folder URLs for every search_docs citation. Live census
over all 58 SUMMARY pages: 47 direct, 10 via redirect, 1 dead.
summaryFolderSlugs now derives the segments from SUMMARY.md; all 60 page
URLs of the real corpus answer 200.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Both variants end with a muted "Learn more" to
troubleshooting/progress-portal-only in the customer docs.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Oliver Kieselbach