Skip to content

Phase 0: correct audit-remediation tail bookkeeping - #166

Merged
ohjonathan merged 1 commit into
mainfrom
codex/audit-tail-bookkeeping
Jul 14, 2026
Merged

Phase 0: correct audit-remediation tail bookkeeping#166
ohjonathan merged 1 commit into
mainfrom
codex/audit-tail-bookkeeping

Conversation

@ohjonathan

Copy link
Copy Markdown
Owner

Summary

Correct the Phase 0 custody record before the remaining audit-remediation sweep:

The live GitHub bookkeeping has also been reconciled:

No product code, tests, workflows, package metadata, tags, releases, or merges are changed.

Baseline and sequencing

  • Baseline: origin/main@3dd093e51e1125147e3533352abda75d7ae1d489
  • Head: 629c984
  • This PR must merge and be independently reverified before PR A starts.
  • Merge, tag, publication, and release actions remain maintainer-owned.

Validation

Passing:

  • scripts/llm-dev doctor --manifest ...
  • llm-dev manifest conformance: 4/4
  • llm-dev scope and cardinality preflights
  • llm-dev staging verification
  • Ontos map/agent regeneration: 209 documents
  • Ontos doctor: 10 passed, 0 failed; only the existing non-Ontos-hook and stale PATH-CLI warnings
  • Ontos link-check: clean
  • git diff --check
  • one complete test-gate run: 1,556 passed

Draft blockers

This PR is intentionally not ready to merge:

  1. Three later full-suite runs reproduced the existing order-dependent SQLite WAL snapshot failure in tests/mcp/test_read_only_registration.py::test_read_only_portfolio_queries_existing_snapshot_without_mutation (1 failed, 1,555 passed); that test passes in isolation. No Phase 0 file is on its execution path, but the program's reliable-full-green guardrail is not yet satisfied.
  2. The user-facing llm-dev lifecycle is honestly review_pending: B.1/B.2/B.3, D.1–D.6, and E remain pending. No external-family receipt or final approval is claimed.

Related issues

@ohjonathan

Copy link
Copy Markdown
Owner Author

Phase 0 review — Claude — approved to merge

Verified the bookkeeping against my independent reconciliation.

SQLite WAL flake — disposition

Provably pre-existing, not a regression from this PR. #166 changed no product or test code, so the code is byte-identical to main — the flake is main's, with certainty (not a "probably pre-existing" guess). It's a test-isolation issue in the portfolio WAL registry (tests/mcp/test_portfolio.py / ontos/mcp/portfolio.py; passes in isolation, fails in some orderings), already tracked as one of the #150 test-hygiene tails under #148.

Not a merge blocker. Fix it as a test-hygiene item during the sweep (naturally alongside PR C). Approved to merge as-is.

One mild note: the full lifecycle machinery (spec + tracker + manifest + session log) is heavier than a ledger/issue bookkeeping change needs — harmless, just noting it.

@ohjonathan
ohjonathan marked this pull request as ready for review July 14, 2026 03:47
@ohjonathan
ohjonathan merged commit bbbad20 into main Jul 14, 2026
6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant