The public repository surface refresh (40d29ca44) hard-wrapped the
Settings → Infrastructure → Install on a host phrase mid-line and
dropped the v5-to-v6 agent upgrades mention from README.md, breaking
RA10's documentation-currentness-tests proof on main.
- documentation_currentness_test.py: match prose fragments against
whitespace-normalized content, the way markdown renders it, so an
editorial rewrap can no longer fail the guard. Structural fragments
(table rows, gate headings) and all assertNotIn checks still match
raw source.
- README.md: restore the v5-to-v6 agent upgrades mention the refresh
dropped; the invariant is still live while v5 installs migrate.
The docs link drift guard (test_repo_docs_link_drift.py) fails CI because
render_release_body.py links docs/CODE_SIGNING_POLICY.md at blob/main.
Release bodies must reference the docs snapshot of the release they
describe, so the installation section now links blob/v{version}, matching
how the rest of the body pins to the version being cut. The unit test pin
follows. Also skip .claude/ in the drift guard so local runs on a shared
tree are not polluted by stale agent worktrees under .claude/worktrees
that CI never checks out.
A `make format` sweep re-lays-out already-committed files without
changing a single token, so it cannot change what renders. The browser
verification guard still demanded a fresh receipt for it, which would
mean recording routes, viewports, states and interactions nobody
exercised in order to describe a diff with no visual delta. A guard that
can only be satisfied by an untrue receipt teaches people to write
untrue receipts.
Exempt a path only when its new content is byte-identical to prettier's
output for its committed content. That is provable, not a judgement
call: if the two match, the sole difference from HEAD is layout.
Fails closed everywhere else -- added or deleted files, unreadable
blobs, prettier missing, or any non-identical output all fall through
and still require the receipt. A reformat that also changes a value is
covered by a test and still blocks.
The pre-commit frontend formatter resolved prettier at
REPO_ROOT/frontend-modern/node_modules/.bin/prettier, where REPO_ROOT is
derived from the script's own path. In a linked worktree that is the
worktree root, which never runs npm install, so prettier_bin() returned
None and the formatter silently returned 0. Every frontend commit made
from a Claude or Codex agent worktree skipped formatting entirely, and
the drift accumulated in already-committed files until someone ran
`make format` and picked up 14 files of unrelated churn.
The skip path's comment claimed "CI's prettier check still catches drift
that slips through here." No such check existed -- nothing under
.github/workflows referenced prettier -- so there was no backstop at all.
- Fall back to the primary worktree's node_modules, resolved via
`git rev-parse --git-common-dir`, and say so in the hook output so a
version mismatch between the two checkouts stays visible.
- Add the whole-tree "Check frontend formatting" step to the frontend CI
job that the comment already promised. Staged-only formatting cannot
see drift in untouched files; this can.
- Pin prettier exactly. "^3.3.0" let the tracked package-lock.json land
on 3.9.5 while the gitignored pnpm-lock.yaml resolved 3.9.6, so CI and
the dev machine were free to run different formatters.
- Resolve the test suite's prettier the same way, so its two real
coverage tests stop silently skipping in worktrees.
Contract-Neutral: devDependency prettier version pin plus pre-commit/CI formatter tooling; no runtime, API, or deployment-surface delta
d00249fd7 added frontend-modern/src/pages/__tests__/Alerts.readOnly.test.tsx
without adding it to the pinned proof list in the canonical completion
guard, so the guard's expected list no longer matched what the registry
derives. canonical-governance has been red on main since that commit, and
the pre-commit hook has been refusing every commit locally for the same
reason. Six commits landed in that window, including the v6.2.0-rc.7 release
preparation.
The delta is one file added and none removed, so this restores the snapshot
rather than lowering the bar. A removal would have meant proof coverage was
dropped, which is a regression to investigate rather than a pin to update.
Contract-Neutral: governance pin restored to match the registry
The "Patrol tools" readiness check read the cached model-readiness
snapshot's tool-protocol dimension on its own. Since the interrupted-run
handling landed (8d0d74e35, b78330405), a run cancelled after every tool
scenario already passed keeps ToolProtocol at pass while the overall
status reports not_assessed, so the check reported "Patrol ready" from an
evaluation that never completed.
The check now requires the snapshot's own overall verdict (Success)
before reporting ready. A snapshot carrying no verdict at all — overall
status not_assessed, or the interrupted or internal_error cause — is not
turned into a failure either: it falls back to the base-config classifier
exactly as an absent snapshot does, capped at a warning. That cap matters
because not_ready is a blocking status in this payload: it clears
readiness.ready, which disables the Patrol run control in
usePatrolIntelligenceState and drops the page into the setup-only view.
#1640 promises a severed or cancelled check never blames the model and
never blocks Patrol from running in Watch mode, and the runtime gate on
POST /api/ai/patrol/run (PatrolRuntimeReadiness) already treats an
unassessed mode as a warning, so a blocking tools check would have
contradicted the route that actually runs Patrol.
A completed run whose tool protocol passed while the overall verdict fell
short now warns instead of claiming ready. It must not block either: the
dimension that actually failed carries the verdict on its own check
(context quality blocks, latency warns), so blocking here would have
turned today's latency warning into a hard stop.
Regression tests: internal/api/issue1640_readiness_gate_test.go covers
the gate across interrupted, internal-error, completed-pass,
completed-fail, and short-of-pass snapshots, asserting the resulting
runnability of the readiness payload;
internal/ai/issue1640_readiness_gate_test.go drives a real evaluation
that is cancelled at the continuation probe to produce the
ToolProtocol=pass / status=not_assessed snapshot end to end and pins that
PatrolRuntimeReadiness keeps Patrol runnable. The new API test file is
registered in the subsystem verification registry.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Two failures on c3fb35c8f. The subsystem lookup unit tests hold their own
hardcoded copies of the settings-shell-and-framing exact_files list, a
third snapshot surface beyond the guard test's, and registering the new
banner proof file left them stale. Synced all eight copies.
The apiClient message-precedence change broke a pre-existing pin that a
short plain-text body outranks the caller fallback on retryable errors
(useReportingPanelState). Body-over-fallback was the long-standing
behavior; what #1640 actually required was dropping markup and oversized
bodies, which stays. Restored body-wins for sane plain text, flipped the
two precedence tests introduced alongside the change, and corrected the
cloud-paid contract paragraph to describe the real order.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Follow-up to 8d0d74e35. The keepalive mechanism was right, the edges
were not.
1. The evaluation ran on a bare goroutine with no recover, so a panic in
provider streaming or validation took the whole Pulse process down.
Before that commit the same panic was on the request goroutine and the
recovery middleware turned it into a logged 500. The goroutine now
recovers, logs the panic with its stack, and answers with an ordinary
readiness result carrying the new internal_error cause and every
dimension reported as not assessed. A Pulse defect is not a model
verdict.
2. Headers were only Set, never committed, despite the comment, the
commit message, and api-contracts.md all claiming otherwise. The
status line went out with the first keepalive at +10s, so a proxy
with a sub-10s time-to-first-byte budget still severed the request.
The transport now writes and flushes WriteHeader(200) before the
ticker starts, matching the pattern the file already uses for SSE.
3. The flusher was resolved with a discarded ok, so a writer that is not
an http.Flusher silently buffered the keepalives and degraded back to
the original bug. It is now checked and logged; the response still
completes, so a warning is the right level here rather than the hard
failure the SSE handlers use.
4. TestIssue1640HandlerUsesKeepaliveTransport grepped the handler source
for substrings, which proves nothing about behaviour. Replaced with a
real httptest.NewServer test that runs a 300ms evaluation and asserts
the client sees the 200 and a body byte before the evaluation
completes, and that the padded body still parses as the expected JSON.
Added coverage for the panic path and the non-flushable writer, and
fixed the eager body[:1] that would panic when a transport regression
left the body empty.
5. The settings readiness banner had no not_assessed branch, so an
interrupted run still rendered the red "Patrol model not verified"
headline: the exact blame-the-model presentation the backend fix
removed. Tone and headline are now exported pure functions with a
neutral treatment for not_assessed and interrupted results, and an
interrupted run cannot claim verification from a max_verified_mode
recorded before the cancellation.
6. createAPIErrorFromResponse let a short plain-text body override an
explicit caller fallbackMessage. A caller passing a fallback knows
which operation it was performing; an intermediary writing the body
does not. Precedence is now canonical JSON, then caller fallback,
then body, with the HTML and oversize suppression unchanged.
7. patrolRunCancelled classified on the raw "context canceled" substring
as its first switch case. Ollama embeds that phrase in its own error
body when it aborts an upstream request, so a genuine provider
failure on a healthy run was classified interrupted and finish()
persisted it as not_assessed. Cancellation is now established from
the run itself (errors.Is(err, context.Canceled), or a cancelled run
context), never from error wording, and the readiness paths classify
through a context-aware entry point. context.DeadlineExceeded keeps
its provider-path timeout classification.
The readiness gate in HandlePatrolModelReadiness keying off ToolProtocol
alone is untouched, as agreed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The registry gained internal/monitoring/issue1638_dns_cache_test.go
(108aa4e20) and internal/api/issue1640_readiness_transport_test.go
(8d0d74e35) as registered verification files, but the expected
verification-requirement snapshots in canonical_completion_guard_test.py
were not updated alongside them, so every Canonical Governance run on
main has been failing its guard unit-test step since. Add both files to
the expected lists.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The alerts-frontend-surface accepted proof list is asserted verbatim by
canonical_completion_guard_test.py, so registering
useAlertOverridesState.test.tsx in registry.json without updating the pinned
expectation failed the guard unit tests.
Verified by running every step of the canonical-governance workflow locally
rather than only the guard I expected to trip: status, control-plane, registry
and contract audits, the Pulse Intelligence gate schema, active-target
automated and hybrid readiness proofs, and all twelve release-control unit
test modules.
rcourtman