After a plan/reflection is elaborated and before it executes, three fresh blind
sub-agents (correctness / robustness / simplicity) attack it on the big model;
the main loop RE-THINKS every aspect a BLOCKER lands (a named plan change, or
[deferred]) and re-challenges once if the plan materially changed. Advisory into
each skill's existing human gate — the human stays the decider.
- lib/challenge-plan.md — reusable phase: fail-safe (never fail open),
severity-driven (any single-lens BLOCKER = must-address), RE-THINK loop
- agents/plan-challenger.md — challenger role (read-only, big-model per BDR-066)
- lib/tests/plan-challenger.test.sh — 41-assertion structure lock
- wired into 11 orchestrators: ship-feature/init-project/feat/bugfix (build-plan),
onboard/audit-delta/code-clean (proposals), seo/geo/harden/web-validate (fix-bundle)
Hardened by dogfooding: 3 blind challengers reviewed this feature's own v1 plan
and caught 4 BLOCKERs (fail-open, consensus-buries-lone-finding, wrong model
tier vs BDR-066, false on-disk-plan premise) — all fixed here.
code-cleaner is now a pure fix executor (was audit+gate+execute). Reroute the two
read-only-audit consumers (onboard STEP 6, tour Phase B) to a big-model agent
(general-purpose/analyzer) — an audit must stay on the big model, never the sonnet
executor. Refactor now runs on sonnet inside the executor (inline-load pin was inert).
Strip the disable-model-invocation frontmatter key from all 19 editable skills. Absent = default = model invocation enabled. 8 were 'true' and blocked the model AND orchestrators from self-routing (status, plugin-check, analyze, onboard, refactor, init-project, pdf-translate, ship-feature) — contradicting the CLAUDE.md skill-routing rules. The other 11 were 'false', a no-op noise line.
The setting is binary (no per-caller granularity), so enabling orchestrator chaining also enables model auto-fire — accepted. Genuinely destructive operations remain guarded by the careful/guard hooks, independent of this flag.
Capitalized: BDR-019 (decision), LRN-026 (learning), journal 2026-06-09.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Before: SEO.md, GEO.md, HARDEN.md, VALIDATE.md, BUGS-FOUND.md landed
at project root. After: all five go to .claude/audits/. Covers both
dispatcher write paths and the dispatcher bash commands that parse
each report (test -s, grep score, wc) — otherwise the dispatcher
would look for the file at the old location.
- skills/seo,geo,harden,validate,code-clean — write paths + console
summaries ("Report: .claude/audits/X.md")
- skills/harden,validate — bash parsing commands (test/grep/wc) aligned
- agents/seo-analyzer,validator-analyzer,code-cleaner — agent-side refs
- agents/validator-analyzer frontmatter description updated
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Skills now delegate to agent .md files instead of embedding logic
inline. Added new agents (bugfixer, code-cleaner, commit-changer,
doc-syncer, feater, hotfixer, seo-analyzer) and new skills
(code-clean, doc, seo). Replaced /readme with /doc (broader scope).
Co-Authored-By: Claude <noreply@anthropic.com>
Bastien Chanot