The 2026-04-22 nightly-dream-cycle learning
(.agents/learnings/2026-04-22-close-loop-citation-gate-deadlock.md) lists
three corrective actions for the chicken-and-egg promotion deadlock that
silently zeroed throughput for weeks:
1. Split admission gates by path — DONE (b69a00f)
2. Parameterize shared helpers (requireCitations bool) — DONE (b69a00f)
3. Fail loudly if ingested > 0 && promoted == 0 — this commit
Prior to this commit, scripts/nightly-dream-cycle.sh reported both numbers
in the summary but never asserted their relationship, so the deadlock
surfaced only in downstream metrics (sigma=0, escape_velocity=false) —
exactly the "quietly posting another zero-throughput status issue" pattern
the learning warns against.
Changes:
- scripts/check-dream-throughput.sh (new): standalone gate consuming
'ao flywheel close-loop --json'. Exits 1 on the deadlock pattern
(added > 0 && promoted == 0), 0 on healthy / no-op, 2 on bad input.
Uses the same pattern as the sibling check-compile-*.sh gates.
- scripts/nightly-dream-cycle.sh: invokes the gate after the summary
artifacts are written (so uploads still succeed) and exits non-zero
when a stall is detected. NIGHTLY_DREAM_STRICT=0 downgrades to warn
for diagnostic runs. The summary.json grows a
close_loop.throughput_stalled bool; the summary.md renders a banner
above the stage table when the stall fires, so the auto-filed
"Nightly dream cycle status" issue leads with it.
- tests/scripts/check-dream-throughput.bats (new): 8 cases covering
healthy, empty-run, deadlock, missing input, no-arg, non-integer
input, and missing keys.
Verified locally against all 8 cases via direct bash invocation
(bats is not available in this environment); shell syntax validated
with 'bash -n' on both scripts.
check-go-absolute-complexity.sh exited 2 with a manual-install hint when
gocyclo was missing, which surfaced go-complexity-ceiling (weight=6) as a
FAIL on any machine whose PATH did not already have gocyclo even though
the repo's code was clean.
Auto-install gocyclo into GOBIN (or $GOPATH/bin) when 'go' is available,
then retry. Provide GOCYCLO_NO_AUTOINSTALL=1 as an opt-out for CI (which
already pre-installs gocyclo in validate.yml) and for offline runs.
Fitness (post-cycle2 -> post-cycle3):
go-complexity-ceiling (w=6): fail -> pass
Verified:
rm /root/go/bin/gocyclo; ao goals measure --goal go-complexity-ceiling
-> auto-install + pass
GOCYCLO_NO_AUTOINSTALL=1 bash scripts/check-go-absolute-complexity.sh
-> exits 2 with install hint (CI-compatible)
https://claude.ai/code/session_015RK27PGX79MsGU5XUhBxkr
compile-freshness and compile-no-oscillation both require .agents/defrag/
latest.json, a local-only artifact that is .gitignored. On machines that
run Dream but not a manual `ao defrag`, both gates fail with "not found"
even though Dream has just written a fresh defrag preview at
.agents/overnight/<run>/defrag/latest.json with the same JSON shape.
- check-compile-health.sh: when the canonical path is missing and
COMPILE_OUTPUT_DIR is unset, fall back to the freshest overnight
preview. CI paths (which set COMPILE_OUTPUT_DIR) remain strict.
- scripts/check-compile-oscillation.sh: new helper that mirrors the
freshness gate's fallback and runs the oscillating-goals jq filter.
- GOALS.md: replace the inline compile-no-oscillation bash with a call to
the new script so the fallback applies there too.
Fitness (post-cycle1 -> post-cycle2):
compile-freshness (w=4): fail -> pass
compile-no-oscillation (w=4): fail -> pass
Verified:
ao goals measure --goal compile-freshness -> pass
ao goals measure --goal compile-no-oscillation -> pass
ao goals validate -> valid: true
heal.sh --strict -> clean
https://claude.ai/code/session_015RK27PGX79MsGU5XUhBxkr
Close the skills/skills-codex ecosystem gaps surfaced by the recent audit:
- docs/SKILLS.md now lists all 69 skills. Added an "Additional Skills" table
covering the 21 previously undocumented entries (autodev, bootstrap, compile,
deps, design, discovery, goals, grafana-platform-dashboard, harvest, perf,
push, readme, red-team, refactor, reverse-engineer-rpi, review, scaffold,
scenario, security, security-suite, test). Removed stale `/knowledge` router
alias and nonexistent `extract` internal-table row.
- Added `scripts/validate.sh` behavioral contracts to the 16 skills that lacked
one (autodev, bootstrap, deps, design, discovery, dream, harvest, llm-wiki,
perf, red-team, refactor, review, scaffold, scenario, test, validation).
Each encodes skill-specific assertions beyond the frontmatter check. Coverage:
53/69 -> 69/69.
- Fixed `scripts/check-codex-parity-drift.sh` which was a silent no-op: it
passed a non-existent --check flag to audit-codex-parity.py, swallowed the
argparse error via `|| true`, then grep'd for strings the audit never emits.
Now uses the exit code + JSON output and correctly fails on drift.
- Added `scripts/check-skill-size.sh` (warn>500, fail>800 lines) to flag
oversized SKILL.md files that should decompose into references/*.md. Current
state: 6 warnings (crank/implement/post-mortem/release/swarm/vibe), 0 fails.
Hygiene sweep: heal.sh --strict clean, codex parity audit clean, 69/69
validators green, size-lint reports no failures.
https://claude.ai/code/session_014AgjT36REb94fjL7sxrnFf
## Docs site — boshu2.github.io/agentops
- Replace Jekyll minima with MkDocs Material (slate dark palette default)
- Landing hero scoped to .landing-hero (no gradient bleed on inner pages)
- Section landings (getting-started, architecture, reference, contracts, levels) with narrative ledes
- Generate skill catalog + per-skill pages from skills/*/SKILL.md at build
- Generate CLI reference from cli/docs/ at build (fails build on missing source)
- Fix every internal link via one-shot migration (scripts/migrations/2026-04-19-fix-docs-links.py)
- Wire mkdocs build --strict into pre-push gate (Check 25a)
- Replace .github/workflows/pages.yml with .github/workflows/docs.yml (actions/deploy-pages)
- README + docs/CONTRIBUTING updated with docs-build toolchain on-ramp
- tests/docs/validate-links.sh patched to understand mkdocs-generated paths
## RPI orchestrator strict-delegation contract
- Strict sub-skill delegation is now the default for /rpi, /discovery, /validation. No --full flag — always on.
- Canonical contract in skills/shared/references/strict-delegation-contract.md (plus Codex mirror)
- Forged learning with detection phrases + corrective actions (ao-inject-ready)
- Rename --no-lifecycle -> --no-scaffold in /discovery (Claude + Codex) with deprecation alias through v2.40.0
- Harden inline-work hotspots with orchestrator-owned markers in /crank STEPS 3a.3, 6.5, 8.7 + Inline Work Policy footer
- Flag consistency: --auto and --interactive scopes documented; --complexity syntax formalized
- CHANGELOG [Unreleased] documents + scaffolds; runtime hook enforcement explicitly deferred
## fix(hooks): go-test-precommit skips when no Go files staged
- Previously ran go vet + go build unconditionally, blocking non-Go commits when Go toolchain was missing or at a different version
- Now checks staged diff and exits cleanly if no *.go or go.{mod,sum,work} files are staged
- Applied to both hooks/ (source) and cli/embedded/hooks/ (synced)
Provenance: full RPI lifecycle with strict delegation at every phase boundary (Phase 1 Skill(discovery) -> Phase 2 Skill(crank) -> Phase 3 Skill(validation) -> Skill(vibe), Skill(post-mortem), Skill(retro)).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Same pattern as cycle 5 (check-goal-staleness.sh): these two scripts were
written against the legacy GOALS.yaml format and now error out (exit 1/2)
when GOALS.yaml is absent, which misleads local runs after the GOALS.md
migration.
- check-pillar-coverage.sh: skip with exit 0 when GOALS.yaml is missing
and GOALS.md is present (pillar taxonomy lives in the narrative now,
not as per-goal pillar: fields the script can machine-check).
- check-goal-quality.sh: skip with exit 0 in the same condition (anti-
pattern detection runs against YAML check: fields that don't exist
in the markdown format).
Both scripts preserve their existing error behavior when neither file
is present — that is still a genuine misconfiguration.
No downstream gate depends on either script today, so this is a
signal-noise fix, not a contract change.
check-goal-staleness.sh was written against the legacy GOALS.yaml
format with machine-readable `added:` and `check:` fields. After
the repo migrated to GOALS.md (see CLAUDE.md §Agent Goals) the script
errors with 'GOALS.yaml not found at GOALS.yaml' — which misleads
readers into thinking a real gate is broken.
When GOALS.yaml is absent AND GOALS.md is present, skip with exit 0
and a clear message that names the migration. Preserve the existing
error behavior when neither file exists (that is still a genuine
misconfiguration). Document the migration in the script header so
the skip path is discoverable.
No downstream gate depends on this script today, so this is a
correctness + signal-noise fix, not a contract change.
The JSONL branch of bootstrap-maturity.sh had two bugs:
1. Detection wrong: `jq -e '.maturity' <file>` only inspects the first
object in a JSONL stream, so files where a later line is missing the
field silently skipped.
2. Output wrong: default `jq` pretty-prints its output, producing
multi-line objects that break JSONL consumers. The unconditional
`. + {"maturity": "provisional"}` also overwrote existing maturity
values (e.g. "stable" → "provisional").
Fix:
- Count missing lines with `jq -c 'select(has("maturity") | not)'`
(per-line stream semantics).
- Transform with `jq -c 'if has("maturity") then . else . +
{"maturity": "provisional"} end'` so existing values are preserved
and output stays valid compact JSONL.
- Add bats coverage for preservation, compact output, and idempotency.
Defines a tracked contract for swarm worker result files living under
.agents/swarm/results/<task>.json. These files were implicitly shaped
but never validated — misshapen packets could silently break crank
result ingestion.
- schemas/swarm-evidence.schema.json: JSON Schema. Required: task,
status, files_changed, notes. Optional: tests, gate, evidence,
before/after counters. additionalProperties=true permits task-specific
extensions.
- scripts/validate-swarm-evidence.sh: jq-based validator accepting a
file or directory. ajv fallback when available. Exit 0 pass, 1 fail.
- docs/contracts/swarm-evidence.md: human-readable contract doc with
example and schema link. Added to docs/INDEX.md catalog.
- scripts/pre-push-gate.sh: new section 26b ('swarm evidence schema')
invokes the validator on every push.
- scripts/check-contract-compatibility.sh: extended ref resolution to
also search schemas/ so docs/contracts/*.md can cite
schemas/*.schema.json without tripping the cataloguing check.
Consumes ag-yhr 'Implement executable swarm evidence schema validation'.
Adds scripts/check-cmd-ao-coverage.sh: reads the existing coverage.out
profile (no second test run), aggregates statement coverage for
cli/cmd/ao/, compares against MIN_COVERAGE floor, exits non-zero on
shortfall. Reads threshold from --min, $MIN_COVERAGE env, or default.
Wires a new 'Enforce cmd/ao coverage floor' step into validate.yml
immediately after the existing go test step, invoking the script with
--profile cli/coverage.out.
Initial floor: 76%. Current measured: 76.8% against the full-suite
coverage.out (2026-04-15). Source epic evolve-cycle-6-coverage-85pct
is driving this to 85%. Ratchet convention: bump MIN_COVERAGE one
whole percent each time real coverage rises, so the floor never
silently regresses.
Consumes evolve-cycle-6-coverage-85pct 'Add 85% coverage gate to CI
validate.yml for cmd/ao' (as a ratchet floor, not a cliff).
scripts/cherry-pick-wave.sh previously removed worktrees unconditionally
in non-interactive contexts and accepted loose branch patterns. Now:
- destructive_consent gate wraps remove_worktrees. Removal fires only
when --yes, --force-delete, or interactive tty detected. Otherwise
prints '[dry-run] would remove worktree: <path>' with instructions.
- Pattern hardening: rejects empty pattern, patterns starting with
wildcards (*, ?, [), and requires first char to match [A-Za-z0-9_-].
- --force-delete flag: explicit opt-in to destructive removal in
non-tty contexts (for CI/automation).
- Help text updated for new flag and safety semantics.
Backward-compat preserved for callers passing --yes (existing test still
passes). New behavior only affects callers running non-interactively
without --yes or --force-delete — they now see dry-run instead of
silent destruction.
Existing test tests/scripts/test-cherry-pick-wave.sh still passes.
shellcheck clean.
Consumes v2.9.3..HEAD 'Harden scripts/cherry-pick-wave.sh pattern
handling and delete safety'.
lib/hook-helpers.sh exports BEADS_DOLT_SERVER_PORT with override via
${BEADS_DOLT_SERVER_PORT:-3307} so users can still override per shell.
All hooks that source hook-helpers.sh automatically inherit the port.
scripts/bd-audit.sh and scripts/bd-cluster.sh are standalone bd callers
that do not source hook-helpers.sh, so they get the same export inline.
cli/embedded/lib/hook-helpers.sh synced via make sync-hooks.
Consumes ag-co7 'Export BEADS_DOLT_SERVER_PORT=3307 in shell env'.
Adds the three missing pieces that turn .agents/ into a proper
LLM wiki following the Karpathy pattern:
1. .agents/INDEX.md — canonical discovery surface cataloging all 202
learnings, plans, research, council reports, and wiki pages with
wikilinks. Auto-generated by scripts/generate-index.sh.
2. .agents/LOG.md — append-only chronological operation log. Entries
formatted as: YYYY-MM-DD HH:MM | actor | VERB | subject | [[wikilink]]
3. .agents/wiki/ directory structure:
- wiki/sources/ — tier 1 LLM session summaries (forge output)
- wiki/synthesis/ — cross-session insights
- wiki/concepts/ — promoted concept pages
- wiki/entities/ — promoted entity pages
Integration:
- ao search now scans wiki/sources/, wiki/synthesis/, wiki/concepts/
(3 new appendKnowledgeMarkdownSearch calls in search.go)
- ao forge --tier=1 writes to .agents/wiki/sources/ (was ao/sessions/)
- Dream post-loop hook + evolve dream sub-cycle write to wiki/sources/
- forge_tier1.go appends to LOG.md and INDEX.md after each session page
- wiki_log.go: AppendToLog + AppendToIndex helpers for all ingest paths
Obsidian-ready: open .agents/ as vault, INDEX.md is the home page,
wikilinks resolve to page navigation, graph view shows connections.
The golangci-lint version mismatch detection was doing an early return
that discarded all findings, even from modules that linted successfully.
In a mixed-version setup (e.g. tests/fixtures/proof-repo uses a different
Go version than cli/), real lint errors from passing modules were silently
dropped and the gate under-reported.
Now the version mismatch is recorded but findings are counted first —
only when zero real findings exist does the tool fall back to "skipped".
https://claude.ai/code/session_01YWC8ZyAyN7kZRXEpNLTN2n
1. Skip TestRunHarvest_PersistsDiscoveryWarnings when running as root
(root bypasses filesystem permission checks, making the chmod(0) trick
ineffective — fixes CLI test failures in root-based CI environments)
2. Detect golangci-lint / Go version mismatch gracefully — when the linter
is built with an older Go than the project targets, treat it as "skipped"
instead of "findings" (fixes false security gate blocks)
3. Fix gosec severity counting in toolchain-validate.sh — previously ALL
findings were added to SECURITY_HIGH_COUNT regardless of severity level,
causing MEDIUM/LOW findings to incorrectly block the security gate. Now
properly categorizes findings by HIGH/MEDIUM/LOW.
Fixes#91, #89, #83https://claude.ai/code/session_01YWC8ZyAyN7kZRXEpNLTN2n
Adds a Windows PowerShell installer for ao, documents the Windows install path, fixes Windows-specific ao/runtime test issues, and removes the CI ripgrep dependency from the Codex override coverage validator.
Single commit covering the tractable slice of the Dream RunLoop Correctness
wake-check follow-ups. na-h61 remains deferred per handoff (requires a
dedicated --complexity=full --mixed session).
ci(validate): fix content-hash gate snapshot wiring (na-lij)
The M9 bracket captured to mktemp but the path was discarded, so the
post-test diff step always hit "ERROR: diff requires a valid snapshot
file path" and the warning fired unconditionally. This explains the M9
"drift on first main run" signal at dbf6d02a. Thread HASH_GATE_SNAPSHOT
through GITHUB_ENV so the diff step actually sees the pre-test path,
and treat a missing snapshot as ::notice:: rather than warn. The gate
now only fires on real drift; ratchet upgrade (warn->fail) is still
gated on 1-2 clean main runs before flipping.
feat(hash-gate): add --ignore-untracked escape hatch (na-5dr)
scripts/check-agents-hash-snapshot.sh: parse --ignore-untracked (no-op
for the ~/.agents/ case since the hub is not a git repo; future-proofs
against a repo'd hub and documents the flag).
scripts/pre-push-gate.sh: honor HASH_GATE_IGNORE_UNTRACKED=1 by
skipping the content-hash block entirely. CI is unchanged (strict
behaviour preserved); operators with local scratch state get a
documented way to push through transient hub drift.
test(overnight): add TestRunLoop_WarnOnlyBudgetExhausted (na-jn9)
Covers the M6 L2 rescue-drain path. Uses an oscillating regression
injector (good<->bad per iter) rather than a plateau injector because
PlateauState.halted is sticky and can only consume one rescue per
run. Regression detection is stateless, so N+1 bad iters drain the
budget exactly. Asserts WarnOnlyBudgetRemaining==0, RegressionReason
contains "warn-only budget exhausted", last iter status ==
StatusHaltedOnRegressionPostCommit, and at least N warn-only rescue
notes in prior iters.
na-khk (bushido env leak in rpi/toolchain_test.go): investigated; zero
"bushido" matches anywhere in cli/, toolchain_test.go has no Setenv/HOME
calls, ./internal/rpi test suite runs clean. Closed as not-reproducible
on HEAD post-M1 HOME isolation (394bfa51).
The existing scripts/check-home-isolation.sh grep-checks test files for
t.Setenv("HOME", ...) patterns but cannot detect:
- Tests that explicitly unset HOME after setup
- Tests that write to ~/.agents and then use os.Chtimes to restore
the mtime (the same API recovery_test.go:285 uses legitimately
can hide test-side poisoning)
This adds a post-hoc content-hash gate that:
1. Captures SHA-256 of ~/.agents/{learnings,patterns,findings,rpi/next-work.jsonl}
BEFORE the test phase.
2. Re-captures and diffs AFTER the test phase.
3. Fails the gate on any content change — mtime restoration cannot
hide it because the hash depends on file content, not stat metadata.
Wired into both local gates:
- scripts/pre-push-gate.sh — capture right after check-home-isolation,
diff as the final phase before the summary.
- scripts/ci-local-release.sh — capture after check_required_cmds,
diff as Phase 6 before Summary via the existing run_step helper.
The remote CI surface (.github/workflows/validate.yml) is deferred to a
follow-up micro-epic so the blast radius of this change stays local.
Script skips gracefully on minimal containers without shasum via an
explicit command-exists check at the top.
Verified via 6-judge pre-mortem council. Part of Micro-epic 1.
Plan rev 2 post-mortem queued 10 follow-up issues; this commit lands
Wave 1 (9 independent issues, parallel workers, zero file conflicts).
Wave 2 (Issue 5 wiring INGEST to real entry points) follows in the
next commit.
Wave 1 issues:
Issue 1 — harvest.WalkOptions.SkipGlobalHub + 2 consumers:
- harvest/walker.go adds the flag, guards the ~/.agents/ auto-include
- corpus/fitness.go opts in so corpus.Compute is hermetic (V8 fix)
- overnight/ingest.go opts in for Dream's private-lane harvest
- 2 new tests: TestDiscoverRigs_SkipGlobalHub_OmitsHome,
TestCompute_Hermetic_DespitePopulatedHome
Issue 2 — forge.RunMinePass in-process entry point:
- NEW cli/internal/forge/runmine.go (272 LOC): pure-extraction
consumer-side entry (mines already-forged sessions, distinct from
cobra's producer-side transcript path).
- cli/cmd/ao/forge.go: thin adapter (+18 LOC); cobra RunE untouched
- 6 new L1 tests in cli/internal/forge/runmine_test.go
Issue 3 — mine.Run in-process entry point:
- cli/internal/mine/mine.go: 134 → 587 LOC (pure extract of git,
agents, code sources; Option B dependency-injection for events
source which depends on cmd/ao internals)
- cli/cmd/ao/mine.go: 580 → 292 LOC (-288), becomes a thin adapter
with type aliases preserving byte-identical cmd/ao shape
- 14 new L1 tests in cli/internal/mine/mine_test.go
Issue 4 — provenance.Audit in-process entry point:
- cli/internal/provenance/provenance.go: +146 LOC adding Audit,
AuditReport, parseLearningFrontmatter, parseLearningDate
- cli/cmd/ao/rpi_phased_provenance.go: thin adapter (+12 LOC)
- 2 new L1 tests
Issue 6 — iterationID sanitization (V9/S1/S2):
- checkpoint.go: sanitizeIterationID helper rejecting /, \, .., empty,
non-printable, whitespace. Called from NewCheckpoint.
- recovery.go: same check in RecoverFromCrash after TrimPrefix extract.
- findings_router.go: findingFilenameRe post-match validated via
time.Parse("2006-01-02") so f-9999-99-99-001.md is rejected.
- 3 new L1 tests
Issue 7 — Stage-order recording hook + test (V6):
- reduce.go: package-level reduceStageRecorder var + call from each
of the 7 REDUCE stages' run closures (nil by default in production).
- stages_test.go: TestRunReduce_FullStageOrder_Enforced asserts the
exact seven-stage order via reflect.DeepEqual.
Issue 8 — Post-commit metadata re-verification (V7):
- checkpoint.go: VerifyMetadataRoundTripPostCommit walks cp.PrevDir
(pre-REDUCE baseline) vs cp.LiveDir (post-Commit live tree),
legitimately-deleted files skipped, strips flagged.
- loop.go: call added AFTER cp.Commit() succeeds, ratchet-forward
(logs a finding, does not unwind). logPostCommitFinding helper
writes to .agents/findings/f-<date>-postcommit-iter-<N>.md
which intentionally does NOT match findingFilenameRe so the
router skips it and it surfaces as a morning-report WARN only.
- overnight_e2e_test.go: TestRunLoop_PostCommitMetadataCheck_LogsFinding
simulates a strip, asserts the finding file is written with the
correct name.
Issue 9 — Package-level TestMain for HOME isolation:
- NEW cli/internal/overnight/overnight_testmain_test.go: TestMain
runs for every test in the package, sets HOME to a temp dir,
pre-creates .agents/learnings so harvest.Promote has a target.
Belt-and-suspenders backup to per-test t.Setenv("HOME", ...) guards.
Issue 10 — check-home-isolation.sh CI gate + pre-push wiring:
- NEW scripts/check-home-isolation.sh: grep-based gate catching Go
test files that reference harvest.*/RunIngest(/harvest.Promote
without t.Setenv("HOME", ...) or a package-level TestMain guard.
Kill switch: CHECK_HOME_ISOLATION_DISABLED=1.
- scripts/pre-push-gate.sh: new check 3b wired after the existing
Go-file section, gated on needs_check go.
- Immediately caught cli/cmd/ao/flywheel_e2e_test.go — fixed inline
with per-test t.Setenv("HOME", t.TempDir()) on all 5 TestFlywheelE2E_*
functions. Defense-in-depth; the tests already used t.TempDir for
promoteDir, so the fix is belt-and-suspenders not a latent bug fix.
Issue 11 — skills-codex parity pass (dream, harvest, evolve):
- skills-codex/dream/SKILL.md: added "Compounding Loop (v2)" section,
3 new $dream start examples, iteration sub-summary output bullet,
"Delineation vs $evolve" table — matching the parent-slice updates
to skills/dream/SKILL.md.
- skills-codex/harvest/SKILL.md: added nightly-usage blockquote.
- skills-codex/evolve/SKILL.md: added "Delineation vs $dream" table
before ## Flags.
- regen-codex-hashes.sh run post-edit to update .agentops-generated
checksums.
Key finding (follow-up bead worth filing):
- scripts/audit-codex-parity.sh reports green even when entire
sections are missing from skills-codex. It catches hash drift
but not semantic content drift. Manual grep was needed to detect
Wave 5 parent-slice updates missing from the codex copies. Tighten
the audit to flag missing headings/sections that exist in
skills/<name>/ but not skills-codex/<name>/.
Gates:
- go build ./... OK
- go vet ./... OK
- go test ./internal/overnight/... ./internal/corpus/... \
./internal/harvest/... ./internal/forge/... \
./internal/mine/... ./internal/provenance/... OK
- go test ./cmd/ao/ (full suite, 183s) OK
- scripts/check-home-isolation.sh PASS
- heal.sh --strict clean
- scripts/regen-codex-hashes.sh 3 skills updated
Parent refs:
- .agents/plans/2026-04-09-dream-followups.md
- .agents/council/2026-04-09-pre-mortem-dream-followups.md
- .agents/plans/2026-04-09-dream-nightly-compounder.md (parent plan)
- .agents/rpi/phase-2-summary-2026-04-09-crank.md (follow-up source)
- .agents/rpi/phase-3-summary-2026-04-09-dream-nightly-compounder.md
1. Add CHANGELOG sync check (check 36) to pre-push-gate.sh — fails if
root CHANGELOG.md differs from docs/CHANGELOG.md
2. Add 250-line limit guard to edit-knowledge-surface.sh for SKILL.md
files — warns at 240, errors at 250
3. Add TestMain to cli/cmd/ao/ that clears AGENTOPS_RPI_RUNTIME* env
vars so per-test t.Setenv is unnecessary
4. Broaden CLI docs parity check (24b) to trigger on any cmd/ao/*.go
change, not just go build changes
5. Wire --fast flag into .githooks/pre-push for faster push-time gates
Bonus: fix pre-existing pipefail bug in flywheel health check (35) that
silently killed all subsequent checks.
- Exclude .gc/ dir and Getenv/DOLT_PASSWORD lines from secret pattern scan
- Regenerate codex hashes for forge, post-mortem, scenario after SKILL.md fixes
Remove coverage-ratchet job, cmd/ao coverage floor check, and all
associated scripts/tests/baseline. Coverage percentages are noisy
during test migrations and architectural refactors — they block CI
without catching real bugs.
Deleted: .coverage-baseline.json, scripts/check-cmdao-coverage-floor.sh,
scripts/coverage-ratchet.sh, tests/scripts/check-cmdao-coverage-floor.bats
Updated: validate.yml, pre-push-gate.sh, ci-local-release.sh, pre-push-gate.bats
Claude