Introduce the core-hub context resolver stabilization_check query, State Hub
progress summaries, activity-definition projections, and unit tests for the
CORE-WP-0007 post-cutover window.
The superseded monolithic event-bridge workplan duplicated T34-T57 task
bindings already tracked in custodian-WP-0003a/b/c. Mark all frontmatter
tasks done, set status archived, and move to workplans/archived/ to clear
fix-consistency C-03/C-15 assessment failures.
Sync the shared issue-core ingestion key from OpenBao into
actcore-runtime-secret via External Secrets, with an interim coulombcore
ClusterSecretStore bootstrap script and deploy docs. Removes manual key
injection from bootstrap-secrets.sh.
Close the intent-gap workplan by documenting the weekly-sbom-staleness
promotion path, adding a null-sink and live REST smoke script, and recording
rollback steps for Railiance. Update SCOPE and deployment docs to reflect
ISSUE_SINK_TYPE=rest and the remaining actcore-runtime-secret key patch.
New context_resolvers/state_hub.py query type: pending_decisions, a thin
GET /decisions/ passthrough (topic_id/workstream_id/workplan_id/decision_type
passed through, status defaults to open). Generic -- not special-cased to
any one decision.
New activity-definitions/monthly-secrets-elevation-review.md: fires 08:00
Berlin on the 1st of each month, sweeps open State Hub decisions under the
infotech/reuse-surface topic, and opens a review task for each. First
target: the temporary autoMode.allow/permissions.allow elevation added to
~/.claude/settings.json on 2026-07-07 (decision 11bf5cbf-458d-4275-a870-
77a82b4058b9, deadline 2026-07-31) for ops-warden/kubectl/OpenBao secret
reads.
Requested by Bernd: no existing scheduling mechanism (session-only cron,
cloud routines with no local access) can durably re-check a local security
posture a month out -- this closes that gap using activity-core's own
durable Temporal-backed trigger instead.
Verified: definition_parser.parse_file + scan_and_parse load it cleanly
alongside the two existing definitions; new resolver tests pass (20/20 in
that file); pending_decisions confirmed against the live local State Hub.
Full suite: 241 passed, 2 pre-existing unrelated failures (confirmed via
git stash -- present before this change too).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Mount the-custodian memory/working from the railiance01 clone (hostPath)
so sweep writeback commits daily-triage notes; worker runs as uid 1000;
progress events store repo-relative working_memory_path; ops inventory
defaults use custodian:// URIs.
Add phase5_stabilization_check State Hub resolver with progress evidence
sinks, schedule projections for daily and closeout checks, custodian:// and
activity-core:// runtime path resolution, and Railiance mounts under /var/custodian.
Unblock lifecycle statuses after daily triage trust evidence closed
(06-28 through 07-03 clean scheduled runs). Finish WP-0010, return
WP-0008/0009/0015 to active with accurate wait reasons, and document
the scheduled-wait vs hard-blocker distinction.
Add a test driving the actual captured 2026-06-26 failure payload
(tests/fixtures/wp0016/...partial.json): it now recovers 6+ valid recommendations
and quarantines the truncated tail, where before WP-0016 it discarded the whole run.
Add an over-depth guardrail test. Together with T03/T04 the regression set now covers
truncation, one-bad-item, oversized-string, over-depth, allow-list/injection-shaped,
and happy-path count cap.
In-repo portion of T05 complete; the live railiance01 graceful-degradation smoke is
operator-owned cluster work (deploy-coupled with the T02 bundle changes) and remains
outstanding. Hand-back notes posted to WP-0006-T03 and WP-0010-T04. Full suite: 220
passed, 1 skipped.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add ADR-004 documenting the producer trust boundary: untrusted producers (LLM,
agent, human; erroneous and malicious), the trust-but-handle vs verify-and-mitigate
postures, error-locality and quarantine-with-provenance principles, and the concrete
activity-core mechanisms.
Implement producer-agnostic guardrails in executor.py, applied uniformly on the
happy path and the recovery path via _partition_items: structural-type -> schema ->
structural caps (_MAX_DEPTH, _MAX_STRING_LEN) -> reference allow-list -> count cap.
Each quarantine carries a reason. Closes the happy-path maxItems count cap deferred
from T03 (valid 9-item report keeps 7, quarantines 2). Reference allow-list reads
context["known_candidates"] via _allow_list_from_context; inert until a resolver
populates it. SCOPE.md updated (executor bullet + ADR list); no INTENT drift.
New tests: happy-path count cap, oversized-string guardrail, allow-list rejection.
Full suite: 218 passed, 1 skipped.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
When the whole-document parse + one retry still fail, report instructions now run
_resilient_report before the total-loss path. A brace/quote-aware scanner
(_extract_object_spans) recovers each recommendation object whether pretty-printed
across many lines or NDJSON one-per-line; a truncated tail gets a best-effort
_try_repair; _partition_items validates each recovered object against the T02 item
schema. Valid items survive (output_validated=True, partial=True), malformed/
over-maxItems items are quarantined with provenance (index, error, raw, reason),
capped at 20. Error locality now matches the unit of work: one bad item costs one
item, not the whole report.
Verified against the real 06-26 shape: 7 valid recommendations + a truncated tail
now recovers all 7 and quarantines the broken tail (previously the whole run was
discarded). Happy-path maxItems top-N enforcement is deferred to T04 (count caps).
Full suite: 215 passed, 1 skipped.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Replace the accept-anything recommendations.items ({type: object}) with a strict
per-item contract (required [rank, candidate, action, why] + typed wsjf) and a
maxItems:7 hint. Strict item structure is what lets the T03 boundary parser
validate each recommendation independently and quarantine only malformed ones.
maxItems is a producer hint (prompt + llm-connect json_schema + T03 mitigation),
NOT a hard reject — a hard maxItems reject would discard a whole 16-item report,
the blast-radius bug WP-0016 removes. DEPLOY COUPLING: the strict schema is also
consumed by the current whole-doc validator, so it must ship with T03's per-item
quarantine parser; until then it increases whole-doc hard-fails. Prompt + max_tokens
headroom + NDJSON framing are documented as a runtime-bundle handoff.
Updated four tests to the strict contract; the forwarded-schema test now reads the
live schema file instead of hard-coding it. Full suite: 213 passed, 1 skipped.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Local analysis of the 2026-06-26 daily-triage validation failure: the unbounded
~1-recommendation-per-workstream list (16 active workstreams; JSON break at char
5268, ~rank 8-9) is the structural cause; both the first attempt and the retry
failed. The exact offending token and finish_reason are unrecoverable from
activity-core data — complete() drops finish_reason/usage, the report sink caps
raw output at 4000 chars (< 5268), and the log preview at 2000. Confirming the
exact token needs llm-connect producer-side logs on railiance01 (operator-owned);
mitigation (T02/T03) is identical regardless. Partial fixture captured.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
