Merge branch 'chore/job9-agents' into develop

2026-07-08 13:23:02 +02:00
parent e4ba8edc16 aa73793b90
commit 02409bbda7
14 changed files with 481 additions and 296 deletions
@@ -80,6 +80,8 @@ rules:
 | BDR-057 | 2026-07-07 | job7: secrets by reference not by value; redact at capture, not just at rest | accepted |
 | BDR-058 | 2026-07-07 | job8: darwin-skill reinstall full pinned tree, detached HEAD (skills CLI single-file-fetch gap) | accepted |
 | BDR-059 | 2026-07-07 | job8: explicit ask-gate for all 4 magic MCP tools, empty allow stays empty | accepted |
 | BDR-060 | 2026-07-08 | job9: CC orchestration floor = v2.1.172 (nested dispatch), supersedes implicit v2.1.83 whole-system floor | accepted |
 | BDR-061 | 2026-07-08 | job9: seo/geo analyzers → fix-bundle→L1 by doctrine (validator-analyzer pattern), not by version constraint | accepted |
 ---
@@ -909,3 +911,23 @@ rules:
 - **Why**: job8 §3/§4 found zero real `mcp__magic__*` invocations ever (transcript census) and one SUSPECT finding (`21st_magic_component_builder` unauthenticated callback-injection channel, [[LRN-110]]). Prior state relied on undocumented absence-means-ask fallthrough — user wants the gate EXPLICIT so it can't silently regress if `permissions.allow` ever gets a careless wildcard or the default-mode semantics change.
 - **Alternatives rejected**: leave everything absent (report's own recommended default) — works today but is silent/undocumented, exactly the posture the user wanted to close; allowlist `logo_search` + `21st_magic_component_inspiration` for frictionless design work (job8 report §3 "frictionless" diff) — explicitly declined, real usage is zero so friction costs nothing.
 - **Reference**: `settings.json` `permissions.ask`, commit `bb7f25a`. Linked to [[LRN-110]] (component_builder risk), [[LRN-111]] (empty-allowlist validity when usage is zero).
 ## BDR-060 — job9: CC orchestration floor = v2.1.172 (nested dispatch), supersedes implicit v2.1.83 whole-system floor
 - **Date**: 2026-07-08
 - **Status**: accepted
 - **Supersedes**: implicit "v2.1.83 = whole-system floor" premise (a misread of [[BDR-004]]'s `decisions.md:133` auto-mode caveat).
 - **Decision**: orchestration floor for any NESTED subagent dispatch = Claude Code **v2.1.172** (nesting stabilized: "let subagents spawn their own subagents", hard cap 5 levels, `Agent` must be in the subagent's `tools:` to nest). Live env confirmed **v2.1.203** (user, nesting supported, cap 5). BDR-004:133 stays UNCHANGED — its `v2.1.83+` is correct for AUTO MODE specifically; the nesting floor is a distinct, higher constraint recorded here (registry is append-only, and BDR-004 is factually right for its scope).
 - **Why**: the whole job1-9 audit series operated on the premise *"CC flattens to 1 level → a 2-level subagent design is silently broken."* That describes the **pre-2.1.172** regime. Corrected in job9 via `claude-code-guide` (official docs `code.claude.com/docs/en/agent-sdk/subagents.md`) + user confirmation of live v2.1.203 → depth findings are VERSION-CONTINGENT, not broken. Path b ([[BDR-061]]) removes the seo/geo analyzers' dependence on nesting, but client-handover's `general-purpose → /seo → seo-analyzer` chain still nests (L1→L2), so the floor stands for the orchestration design.
 - **Alternatives rejected**: keep the implicit v2.1.83 floor — predates nesting, mislabels version-contingent flows as "BROKEN"; hard-gate CC version in `doctor.sh` — deferred (path b de-risks the analyzers; a doctor warn-gate is an optional follow-up, and `doctor.sh` is config-guarded → sentinel cost not justified now); raise BDR-004:133 to v2.1.172 — WRONG, that caveat is auto-mode-specific (auto mode works from 2.1.83) and rewriting it would violate append-only + inject a factual error.
 - **Reference**: `.audit/job9-report.md` §Premise + §6 D-version-floor; `decisions.md:133` (BDR-004 auto-mode caveat, unchanged). Linked to [[BDR-061]] (path-b), [[LRN-112]] (nesting mechanics).
 ## BDR-061 — job9: seo/geo analyzers emit a fix-bundle applied at L1 by doctrine (validator-analyzer pattern)
 - **Date**: 2026-07-08
 - **Status**: accepted
 - **Decision**: `seo-analyzer` + `geo-analyzer` re-architected to the `validator-analyzer` contract — they AUDIT and EMIT a machine-parseable `## FIX BUNDLE` terminated by the verbatim `READY TO APPLY — awaiting dispatcher confirmation` sentinel; they NEVER edit code and NEVER dispatch a sub-agent (`Agent` dropped from both `tools:`). The DISPATCHER applies at **L1 from its own main loop**: `/seo` (new STEP 1.5) + `/geo` (rewritten to dispatch+apply, mirrors `/web-validate`) dispatch `hotfixer`/`feater` at L1; `/harden` keeps its existing direct-Edit STEP 3 (already end-to-end path-b); `/onboard` stays audit-only (bundle produced, deferred to backlog STEP 9). AUTO tier applies unconfirmed; GATED tier (seo D/E · geo G5) requires explicit accord; USER ACTIONS → report §11.
 - **Why**: by DOCTRINE, not version constraint. Before: analyzer STEP 12/13 dispatched hotfixer/feater; when the analyzer was itself a subagent (`/seo` → analyzer at L1), that dispatch was **L2 nesting** → silent no-op on CC<2.1.172, and both analyzers forbade direct edits → the reported bug: *report produced, ZERO fix applied*. The bundle→L1 pattern (a) lands fixes on ANY CC version (single dispatch level), (b) gives fresh-context specialist fixes without depth risk, (c) dissolves the `/seo` parallel-edit race (fixes now applied serially by the dispatcher, by file ownership). `/harden` already proved the pattern in-repo. Chosen even though [[BDR-060]] confirms live nesting works — version-robust by design beats version-contingent.
 - **Alternatives rejected**: only raise the version floor (BDR-060 alone) — leaves the analyzers version-contingent, and the `/seo` nested-fix design fragile; keep analyzers self-applying but require CC≥2.1.172 — works on current env but not robust and keeps the parallel-edit race; make the dispatcher apply via direct Edit everywhere (like /harden) instead of hotfixer/feater — loses the fresh-context specialist fix; kept direct-Edit only for /harden's tiny scope.
 - **Verification**: `make test` green + 4 real smokes — analyzer emits bundle + edits nothing (md5 unchanged); AUTO fix lands on disk via L1 hotfixer with no confirmation (the exact previously-broken path); GATED withheld pre-approval then applied post-accord; /onboard writes only the report, zero source files.
 - **Reference**: `agents/seo-analyzer.md` STEP 12, `agents/geo-analyzer.md` STEP 13, `skills/seo/SKILL.md` STEP 1.5, `skills/geo/SKILL.md`, `agents/validator-analyzer.md` (reference contract), `.audit/job9-report.md` §6 option (b); commits `a5a7b54`/`6df42e4`/`c498b93`/`70fb3b4`. Linked to [[BDR-060]] (nesting floor), [[LRN-112]] (nesting mechanics).
@@ -358,3 +358,10 @@ rules:
 - job8 third-party security audit shipped read-only: `.audit/job8-report.md` — magic MCP/plugins/gstack/external skills/trust chain, 9 explorers + verifier batches, 11 CONFIRMED/5 CORRECTED/0 REFUTED. Surfaces C (ui-ux-pro-max) + D (other plugins) finished inline, single-observer, no verifier pass — Fable-5 spend limit hit mid-run.
 - User GO on all 4 items: A allowlist stays empty, ask-gate explicit; B covered by A (no STOP); C reinstall pinned (not remove/keep-broken); D no action. Executor = this session, `chore/job8-hardening`, no finish.
 - job8 EXECUTED: 3 commits. **A**: `settings.json` `permissions.ask` += 4 `mcp__magic__*` tools, isolated from 2 unrelated pre-existing edits (model/skipWorkflowUsageWarning) already sitting uncommitted before this session started — those restored uncommitted after, not part of this branch's history [[BDR-059]]. **B**: confirmed `component_builder` in scope of A's gate, no STOP needed; documented the callback-injection risk in README's MCP section + [[LRN-110]] — third-party package code, not patched. **C**: confirmed referenced files (`references/`, `scripts/`, `templates/`) 100% absent from `~/.agents/skills/darwin-skill/` (only `SKILL.md` present) — root-caused to the `skills` CLI's `skillPath` install field fetching a single file, not the repo tree [[LRN-109]]. Upstream HEAD matched the already-recorded lockfile hash exactly (zero drift). Reinstalled full tree at that pinned SHA, `.git` kept but detached (2nd real SHA-pin after gstack) [[BDR-058]]. Backup of old single-file dir kept. Git-commit whole-`.claude/skills`-tree scope NOT restricted (3rd-party pinned code, patching breaks the pin) — documented as accepted risk instead. 3 Bash permission denials mid-C (rsync x2, cp+rm) before a plain `cp` succeeded — `rm -r*`/`rm -rf*` are hard-denied even for scratch/temp paths, no prompt possible; switched approach rather than retrying identically. **D**: confirmed untouched. `make test` green throughout (incl. a live `path_present(darwin-skill)` fs check). Smoke gate: real `mcp__magic__logo_search` call in-session, user confirmed the ask prompt fired and was manually approved — no auto-exec. [[LRN-111]]. Branch unmerged, human gate. **Not re-verified this cycle** (job8 report's own caveat, carried forward): surfaces C/D (ui-ux-pro-max, other plugins) were single-observer CLEAN findings with no adversarial pass — re-audit next cycle if darwin/magic scope comes up again.
 ## 2026-07-08
 - job9 sub-agent architecture corrections shipped, `chore/job9-agents`, 10 code commits, `make test` green throughout. Premise correction confirmed: CC **v2.1.203** live, nesting supported (cap 5, `Agent`-in-tools required) — [[LRN-112]], contradicts the operating premise of the whole job1-9 series.
 - **Part 1** (4 commits, `0ede52c`..`5ab6c21`): commit-changer drop unused `Agent`; verifier + security-auditor + plugin-advisor pinned `model: sonnet`. Gate = real dispatch smoke on sonnet: verifier `CONFORME`, security-auditor `BLOCK(2)` (checklist caught planted hardcoded-secret + SQLi that semgrep 1.168.0 missed), plugin-advisor `ACTION REQUIRED` — verdict grammar intact, mode honored, no revert.
 - **Part 2** (`a5a7b54`/`6df42e4`/`c498b93`/`70fb3b4` + hardening `212f9aa`): seo/geo analyzers re-architected to fix-bundle→L1 (validator-analyzer contract), `Agent` dropped from both `tools:`; `/seo` new STEP 1.5 applies at L1 (serial by ownership, dissolves the parallel-edit race), `/geo` → dispatch+apply orchestrator, `/harden` already end-to-end path-b (untouched), `/onboard` audit-only (untouched). [[BDR-060]] version floor + [[BDR-061]] path-b doctrine. 4 real smokes green: analyzer emits bundle + edits nothing (md5 unchanged, no files created); AUTO fix LANDS on disk via L1 hotfixer with no confirmation (the exact previously-broken path — *report but zero fix* → resolved); GATED withheld pre-accord then applied post-accord (new tier, first test); /onboard writes only the report, zero source files.
 - **Part 3** (`87d63bf`/`af9656f`): H2 "Load and follow" idiom → **INLINE-LOAD** verb at code-cleaner + scaffolder (main-loop-BECOMES-agent, `Agent` not involved), drop unused `Agent` from code-cleaner; H1 code-cleaner→refactorer handoff now a named artifact `.claude/audits/CODE-CLEAN-SCOPE.md`. Tight scope per user (2 cited sites, no 40-site rewrite).
 - Branch unmerged, human gate. **Fixed** (`5a3de92`, isolated): stripped `Co-Authored-By: Claude` from `commit-changer.md` message template — it contradicted [[no-commit-attribution]] since the template's creation (the settings.json backstop caught real commits, but the template itself would keep re-seeding the trailer). Only banned trailer in the file (no Claude-Session/--trailer). FOLLOW-UP next cycle: cross with J4-16 (lib-layer lock) to verify no other agent template carries the same trailer.
@@ -125,6 +125,7 @@ rules:
 | LRN-109 | 2026-07-07 | job8: `skills` CLI (vercel-labs/skills) fetches only `skillPath` (often just SKILL.md), not sibling refs/scripts/templates the skill text references — darwin-skill install gap, not drift/tamper | installing/auditing any skill via the `skills` CLI whose SKILL.md references relative paths — verify those paths exist post-install, don't trust `skillFolderHash` alone |
 | LRN-110 | 2026-07-07 | job8: `21st_magic_component_builder` (magic MCP) opens unauth'd 127.0.0.1 callback server, CORS `*`, no token check, 10min window — any local POST lands verbatim in the tool result the model consumes = local prompt-injection channel | any MCP tool that opens a local callback/listener server to receive async results — check auth + origin scoping on the listener, not just the outbound call |
 | LRN-111 | 2026-07-07 | job8: empty permissions.allow for a risky MCP tool is a VALID posture (not a gap) when transcript census shows zero real invocations — pre-authorizing unused surface buys nothing, ask-gate costs nothing | deciding whether to allowlist any tool/command — check real usage before assuming "no entry = todo" |
 | LRN-112 | 2026-07-08 | job9: CC nested subagent dispatch SUPPORTED since v2.1.172 (cap 5 levels, `Agent` must be in subagent `tools:`) — "flattens to 1 level" is the pre-2.1.172 regime; live env v2.1.203. Contradicts the operating premise of the whole job1-9 series | a subagent-dispatches-subagent design is VERSION-CONTINGENT, not "broken" — check CC version before flagging; fix = raise floor or re-architect to bundle→L1 |
 ---
@@ -1123,3 +1124,9 @@ rules:
 - **context**: job8 census (grepping real `"name":"mcp__…"` tool_use blocks across `~/.claude/projects`, not text mentions) found ~910 mentions of `mcp__magic__*` but ZERO real invocations, ever. `permissions.allow`/`permissions.ask` had no `mcp__*` entries at all before this job — job6 flagged that as "ZERO scoping", easy to misread as an oversight to fix by adding an allowlist.
 - **future application**: before treating "no entry for tool X" as a gap needing an allowlist, check real usage first (grep tool_use blocks, not prose mentions). If usage is zero, pre-authorizing costs nothing to skip and buys nothing to add — the honest fix is making the ask-gate EXPLICIT (so it can't regress silently), not granting allow access nobody needs yet. Only add allow entries when real, measured, recurring usage justifies removing the friction.
 - **cousin**: [[BDR-059]], [[LRN-110]], [[LRN-088]] (same family: measure before assuming an absence is a defect).
 ## LRN-112 — nested subagent dispatch is supported (CC ≥ v2.1.172), not a flatten-to-1 no-op
 - **context**: the whole job1-9 audit series ran on the premise *"Claude Code aplatit à 1 niveau → un design supposant 2 niveaux de sous-agents est cassé silencieusement."* job9 corrected it via `claude-code-guide` (official docs `code.claude.com/docs/en/agent-sdk/subagents.md`): a running subagent CAN spawn a further subagent IF `Agent` is in its `tools:` (omit it / add to `disallowedTools` to prevent nesting); hard cap **5 levels** ("a subagent 5 levels below main can't spawn further"); nesting **stabilized in v2.1.172** ("let subagents spawn their own subagents") — earlier versions did not support it at all. Live env confirmed **v2.1.203** (user). `claude --version` was unavailable in-sandbox so the report bracketed but could not pin it; the user pinned it.
 - **future application**: NEVER classify a subagent-dispatches-subagent design as "BROKEN" without checking the CC version. On ≥2.1.172 it works within the 5-level cap; on <2.1.172 it silently no-ops. The actionable finding is a VERSION-FLOOR ([[BDR-060]]) or a version-robust re-architecture (bundle→L1, [[BDR-061]]) — not "it's broken." When an agent must NOT nest, enforce it structurally: drop `Agent` from its `tools:` (done for seo/geo analyzers). Re-audit any prior job1-9 "nested = broken" finding through this lens.
 - **cousin**: [[BDR-060]] (version floor), [[BDR-061]] (path-b bundle pattern), [[LRN-057]] (subagent invocation idioms).
@@ -1,5 +1,42 @@
 # TODO
 ## 2026-07-08 — job9 sub-agent architecture corrections (chore/job9-agents)
 Genèse : `.audit/job9-report.md` (agents/*.md frontmatter+body, verify-loop,
 dispatch graph, read-only). Premise correction confirmed CC v2.1.203 : nesting
 SUPPORTED since v2.1.172, cap 5, `Agent` tool required in `tools:` to nest.
 User decision: **path b (version-robust)** for the version-floor. One commit/item.
 PART 1 — MISROUTED (trivial frontmatter):
 - [x] A — commit-changer: drop unused `Agent` from tools (0ede52c)
 - [x] B — verifier: pin `model: sonnet` (ea6c126)
 - [x] C — security-auditor: pin `model: sonnet` (1c270e6)
 - [x] D — plugin-advisor: `haiku` → `sonnet` (5ab6c21)
 - [x] GATE P1 — smoke green: verifier CONFORME, sec-auditor BLOCK(2), advisor
      ACTION REQUIRED; verdict grammar intact, mode honored. No revert.
 PART 2 — VERSION-FLOOR (path b) — CONTRACT APPROVED, DONE:
 - [x] 5 — seo+geo analyzers → fix-bundle→L1 (a5a7b54/6df42e4); /seo STEP 1.5
      (c498b93), /geo dispatch+apply (70fb3b4), dispatcher tier-tolerance
      (212f9aa); /harden already path-b (untouched), /onboard audit-only
      (untouched). GATE PASSED: make test green + 4 smokes (A bundle-no-edit,
      B AUTO lands on disk no-confirm, C GATED withheld→applied post-accord,
      D onboard report-only zero-fix).
 - [x] 6 — BDR-060 orchestration floor v2.1.172 supersedes implicit v2.1.83
      premise (BDR-004:133 kept — auto-mode floor, append-only + factually
      correct). BDR-061 path-b doctrine.
 PART 3 — IMPLICIT-HANDOFF (tight scope, 2 sites) — DONE:
 - [x] 7 — H2 INLINE-LOAD verb @ code-cleaner + scaffolder (87d63bf/af9656f),
      drop unused Agent from code-cleaner
 - [x] 8 — H1 code-cleaner→refactorer named artifact .claude/audits/CODE-CLEAN-SCOPE.md
 Capitalize DONE: LRN-112 (nesting) + BDR-060 (floor) + BDR-061 (path-b) + journal.
 - [x] commit-changer template Co-Authored-By stripped (5a3de92, isolated) —
      contradicted no-attribution ban since creation
 - [ ] FOLLOW-UP next cycle: cross with J4-16 (lib-layer lock) — verify no other
      agent/template carries a banned attribution trailer (Co-Authored-By/
      Claude-Session/--trailer)
 Branch unmerged, human gate.
 ## 2026-07-07 — job8 third-party security hardening (chore/job8-hardening)
 Genèse : `.audit/job8-report.md` (magic MCP/plugins/gstack/external skills/trust
 chain, read-only). A/B/C/D exécutés (3 commits), branche non mergée, gate humain.
@@ -1,7 +1,7 @@
 ---
 name: code-cleaner
 description: Audit codebase for dead code, style violations, and structural issues. Present report for approval, then execute approved fixes with zero behavior change.
-tools: Read, Edit, Write, Bash, Grep, Glob, Agent, AskUserQuestion
+tools: Read, Edit, Write, Bash, Grep, Glob, AskUserQuestion
 ---
 # CODE-CLEAN — Codebase Cleanup
@@ -128,14 +128,24 @@ and ask for explicit per-item confirmation.
 ### STEP 5 — STYLE FIXES + STRUCTURAL REFACTORING
-For approved style and structural items:
+For approved style and structural items, hand off to the refactorer:
-1. Load and follow `$HOME/.claude/agents/refactorer.md`
+1. **Persist the handoff contract.** Write the approved items to
-2. Pass the approved list as the refactoring scope
+   `.claude/audits/CODE-CLEAN-SCOPE.md` (run `mkdir -p .claude/audits`
-3. The refactorer handles the actual code changes with its own
+   first), one per line in the report format `file:line — item —
-   safety process (pre-report, function-by-function, test after each)
+   severity — proposed fix`. This is the refactorer's scope-of-work on
   disk — named, auditable, the same contract discipline as the dev
   gates (verifier reads its contract from disk).
 2. **INLINE-LOAD the refactorer.** Load `$HOME/.claude/agents/refactorer.md`
   and continue AS the refactorer in THIS SAME context — you *become* it.
   This is an inline load, NOT a subagent dispatch: the `Agent` tool is
   not involved and no new context is spawned. Its scope = the items in
   `.claude/audits/CODE-CLEAN-SCOPE.md`.
 3. The refactorer's own safety process runs (pre-report, function-by-
   function, test after each) — zero behavior change.
-Do NOT call the `/refactor` skill — invoke the agent directly.
+Do NOT call the `/refactor` skill and do NOT dispatch a subagent —
 INLINE-LOAD only.
 ### STEP 6 — LOG DISCOVERED BUGS
@@ -1,7 +1,7 @@
 ---
 name: commit-changer
 description: Retrace-and-commit engine — dispatched by /commit-change. Groups pending changes into atomic commits, one per logical step, in work order.
-tools: Bash, Read, Grep, Glob, Agent, AskUserQuestion
+tools: Bash, Read, Grep, Glob, AskUserQuestion
 ---
 # Git Smart Commit
@@ -105,8 +105,6 @@ Follow Conventional Commits and match the repo's existing style:
 <type>(<scope>): <short description>
 <optional body — what and why, not how>
 Co-Authored-By: Claude <noreply@anthropic.com>
 ```
 Types: `feat`, `fix`, `refactor`, `chore`, `docs`, `test`, `style`, `perf`
@@ -1,7 +1,7 @@
 ---
 name: geo-analyzer
-description: GEO audit agent for AI search engines — dispatched by /geo and /seo. Audits AI crawlers, llms.txt, entity signals, Schema.org; autonomous fixes, scored report. Classical SEO → seo-analyzer agent.
+description: GEO audit agent for AI search engines — dispatched by /geo and /seo. Audits AI crawlers, llms.txt, entity signals, Schema.org; emits a fix bundle (dispatcher applies), scored report. Classical SEO → seo-analyzer agent.
-tools: Read, Edit, Write, Bash, Grep, Glob, Agent, WebFetch, WebSearch
+tools: Read, Edit, Write, Bash, Grep, Glob, WebFetch, WebSearch
 ---
 # GEO — Generative Engine Optimization audit, fix & strategy
@@ -595,7 +595,7 @@ High-impact, low-effort. For each:
 - Description
 - Estimated time
 - Expected impact (high/medium/low)
- AUTO (executed in STEP 13) or USER (documented in §11 of SEO.md)
+- AUTO (bundled in STEP 13, applied by the dispatcher) or USER (documented in §11 of SEO.md)
 **MANDATORY user action — AI index submission**: every FULL audit
 MUST emit these 3 user actions (they are the entry points for AI
@@ -643,119 +643,90 @@ Consolidate EVERY finding from STEPs 4-9 into structured batches.
 | **G6 — Entity @id + sameAs wiring** | `feater` | JSON-LD graph restructure | No |
 | **G7 — User actions** | documented in §11 | Wikidata, KP, monitoring | N/A |
-Print the plan before STEP 13.
+Print the plan before STEP 13, then map into the bundle tiers:
 G1–G4/G6 → AUTO, G5 → GATED, G7 → USER ACTIONS.
-**User unreachable / headless run → ALL batches become report-only,
+**Apply-vs-report is the DISPATCHER's call, not yours.** You ALWAYS emit
-including the "Confirmation: No" ones.** Autonomous batches presume a
+the bundle (STEP 13) and NEVER apply — you neither edit nor create files
-reachable user who saw the printed plan and can interrupt. With nobody
+(robots.txt, llms.txt, JSON-LD) under any condition. The dispatcher decides
-watching, modify NOTHING: document every proposed fix in the report
+whether to apply it (reachable user / auto flow like /seo, /geo) or leave
-(§9/§11) with its ready-to-apply content, and leave source files,
+it as a report (headless/CI run, or an audit-only flow like /onboard). This
-robots.txt and llms.txt untouched/uncreated. Next reachable run applies
+removes the old analyzer-side "reachable?" branch — the decision now lives
-them after the plan gate.
+one level up, where the plan is printed and the user can interrupt.
 Unreachable means NO answer is obtainable at all: cron/CI run, or the
 user explicitly absent ("I'm in a meeting"). Being dispatched as a
 subagent by an orchestrator (e.g. /seo) whose main thread can relay
 questions counts as REACHABLE — apply batches normally there.
 ---
-## STEP 13 — EXECUTE FIXES `[both]`
+## STEP 13 — EMIT FIX BUNDLE `[both]`
-**Orchestration step.** Delegate to specialist agents. Do NOT edit
+**You do NOT apply fixes and you do NOT dispatch any sub-agent.** Same
-files directly.
+contract as `validator-analyzer` and `seo-analyzer`: serialize the STEP 12
 batches into a machine-parseable FIX BUNDLE. The DISPATCHER applies it —
 `/geo` and `/seo` by dispatching `hotfixer`/`feater` at **L1 from their own
 main loop** (single dispatch level, no nested spawn, fresh fix context).
 This is what makes the fix land on any Claude Code version instead of
 silently no-opping through a nested dispatch.
-### G1 — robots.txt AI directives
+Tier mapping: G1–G4/G6 → AUTO, G5 → GATED, G7 → USER ACTIONS.
 ### Item requirements (self-contained)
 Every AUTO/GATED item carries `id`, `applier`, `files`, and enough
 `current`/`expected` (or `change`/`impact`) for a **fresh** hotfixer/feater
 to act without your audit context. Embed per item:
 - **Shared-file edit discipline** — on shared templates (Layout.astro,
  index.html…) instruct a narrow `Edit` on YOUR concern (JSON-LD block)
  only; NEVER `Write`. `Write` only on sole-owned files (robots.txt,
  llms.txt, llms-full.txt).
 - **Templates + context** — G2/G6 paste the expected JSON-LD from
  `geo-schemas.md` + business context (entity name, sameAs, @id canonical)
  + framework note. G4 follows `llms-txt-template.md` exactly. G1 pastes
  the correct variant from `ai-crawlers-2026.md`.
 - **PERMISSIVE default** on G1 unless the client flagged premium/regulated.
 ### Output shape
 Spawn `hotfixer`:
 ```
-SEO/GEO hotfix: update robots.txt to <PERMISSIVE|RESTRICTIVE> AI crawler strategy.
+## FIX BUNDLE (for dispatcher)
-File: robots.txt
+
-Current state: <list directives present + missing>
+### AUTO — apply without confirmation
-Expected state: <paste from ai-crawlers-2026.md, correct variant>
+- id: G1
-Context: GEO audit, autonomous scope. No confirmation needed.
+  applier: hotfixer
  files: robots.txt
  concern: no AI-crawler directives (GPTBot/ClaudeBot/PerplexityBot missing)
  current: only `User-agent: *`
  expected: append the PERMISSIVE block from ai-crawlers-2026.md (Write — sole owner)
 - id: G2
  applier: hotfixer
  files: src/layouts/Base.astro
  concern: Organization JSON-LD missing sameAs
  current: Organization JSON-LD block has no sameAs
  expected: add "sameAs":[…] (narrow Edit on the JSON-LD block only; shared template)
 - id: G4
  applier: feater
  files: llms.txt (new) + build generator
  concern: llms.txt absent (GET /llms.txt → 404)
  current: no file
  expected: create per llms-txt-template.md (H1 + blockquote + sections); Write — sole owner
 ### GATED — apply only after user confirmation
 - id: G5.1
  applier: feater
  files: src/pages/index.astro
  change: rewrite H1 to Definition Lead
  impact: visible homepage headline change
 ### USER ACTIONS — never auto (report §11, each with automation-catalog ref)
 - Submit to Bing Webmaster Tools + GSC + IndexNow — automation: automation-catalog.md
 - Wikidata entity creation — automation: <catalog ref>
 READY TO APPLY — awaiting dispatcher confirmation
 ```
-### G2 — Schema.org fixes (parallel if independent files)
+Emit the `READY TO APPLY — awaiting dispatcher confirmation` line
-
+**verbatim** as the bundle's last line — the dispatcher keys its apply step
-Spawn `hotfixer` per file OR `feater` if cross-file graph restructure.
+on it. Do NOT run JSON-LD/robots.txt/llms.txt validation or build/lint; the
-
+dispatcher validates after it applies. Your job ends at the sentinel.
 Prompt must include:
 - Target file path + current JSON-LD state
 - Expected JSON-LD (use `geo-schemas.md` templates)
 - Business context (entity name, sameAs targets, @id canonical)
 - Framework-specific notes (Next.js metadata export, Astro component props, etc.)
 ### G3 — Remove deprecated schemas
 Fast `hotfixer` pass. One per file or one consolidated.
 ### G4 — llms.txt creation
 Spawn `feater`:
 ```
 GEO feature: generate llms.txt (and llms-full.txt if documentation site).
 Files to create: /llms.txt + endpoint/generator to rebuild on deploy.
 Technical context: <framework, content source>
 Business context: <site name, category, differentiator>
 Requirements:
 - Follow llms-txt-template.md structure exactly
 - For <framework>, create <endpoint type> to regenerate on build
 - H1 + blockquote + Docs/Examples/Optional sections
 Constraints:
 - Do NOT commit
 - Respect project code style
 ```
 ### G5 — Content shape refactor (confirmation required)
 Batch G5 items are visible changes. Present full list to user:
 ```
 CONTENT SHAPE CHANGES — approval needed:
  G5.1 Homepage H1 — change from "<current>" to Definition Lead "<new>"
  G5.2 /services page — add TL;DR block
  G5.3 Blog template — move summary above fold
  ...
 Approve all / select / skip?
 ```
 For approved: spawn `feater` with detailed spec.
 Unapproved → document in §9 (medium term) of SEO.md.
 ### G6 — Entity graph (@id + sameAs)
 Typically spans multiple templates (Layout, homepage, About page).
 Single `feater` call with full restructure spec.
 ### G7 — User actions
 Document in SEO.md §11. No execution. Every entry MUST include
 "Automatisation possible avec: ..." per `automation-catalog.md`.
 ### Verification
 After all sub-agents complete:
 1. **Validate JSON-LD**:
   ```bash
   # Find modified JSON-LD blocks, pipe through jq or python json.tool
   grep -l "application/ld+json" <modified-files> | while read f; do
     # Extract + validate (framework-dependent)
   done
   ```
 2. **Validate robots.txt**:
   ```bash
   # No duplicate User-agent directives? No Disallow without User-agent?
   [ -f robots.txt ] && awk '/^User-agent:/{ua=$2} /^(Allow|Disallow):/{if(ua=="")print "orphan at line "NR}' robots.txt
   ```
 3. **llms.txt shape**:
   ```bash
   [ -f llms.txt ] && head -1 llms.txt | grep -q "^# " && sed -n '2,10p' llms.txt | grep -q "^> " && echo "llms.txt header OK"
   ```
 4. **Build/lint if available**: `npm run build`, `npm run lint`.
 Revert any sub-agent change that breaks build.
 ---
@@ -797,8 +768,11 @@ without evidence = DGCCRF risk.>
 <Each entry MUST include "Automatisation possible avec:" per
 automation-catalog.md>
-## ENTRIES FOR SEO.md §15 (change log):
+## ENTRIES FOR SEO.md §15 (change log — filled by the DISPATCHER after it applies the bundle):
-<Every file modified, what was changed, why, verification status>
+
 ## FIX BUNDLE (for dispatcher):
 <the AUTO / GATED / USER ACTIONS block from STEP 13, ending with the
 verbatim `READY TO APPLY — awaiting dispatcher confirmation` sentinel>
 ## GEO SCORING:
 <Axes scoring block from STEP 10>
@@ -866,10 +840,13 @@ PROCHAINE ETAPE : <highest-priority>
 ## RULES
 ### Orchestration
- **Analyze before fixing.** STEPs 0-12 are pure analysis. No file
+- **Analyze, then bundle — never apply.** STEPs 0-12 are analysis;
-  modification until STEP 13.
+  STEP 13 emits a FIX BUNDLE. You NEVER edit a code file (report files
- **Delegate.** Never edit JSON-LD / robots.txt / llms.txt directly
+  only) and NEVER dispatch a sub-agent — the dispatcher applies the
-  in STEP 13. Use `hotfixer`/`feater` with self-contained prompts.
+  bundle at L1 (single dispatch level, lands on any Claude Code version).
 - **Bundle items are self-contained.** Each carries file paths, current
  vs expected JSON-LD/robots.txt/llms.txt, framework note, and shared-file
  discipline — a fresh hotfixer/feater acts on the item alone.
 - **Depth-aware.** LOCAL skips STEPs 3, 9. Same rigor elsewhere.
 - **Standalone vs dispatched.** If dispatched via `/seo`, output the
  structured envelope in STEP 14. Standalone (`/geo`), write GEO.md
@@ -881,8 +858,9 @@ PROCHAINE ETAPE : <highest-priority>
  duplicate. Reference them in §13 as "see SEO section" if needed.
 - **Shared-file edit discipline.** On template files shared with
  `seo-analyzer` (Layout.astro, index.html, base.html.twig, etc.),
-  your sub-agents (`hotfixer`/`feater`) MUST use `Edit` with a narrow
+  each bundle item MUST instruct the applier (`hotfixer`/`feater`) to
-  `old_string` targeting ONLY your owned concern (JSON-LD block).
+  use `Edit` with a narrow `old_string` targeting ONLY your owned
  concern (JSON-LD block).
  NEVER `Write` on shared templates. `Write` is reserved for files
  you solely own: robots.txt, llms.txt, llms-full.txt. Full-template
  refactor → escalate as user action in §11.
@@ -905,6 +883,6 @@ PROCHAINE ETAPE : <highest-priority>
  `automation-catalog.md`. No exceptions.
 - **WebSearch on FULL audits** to cross-check crawler list + tool
  landscape before emitting — these shift quickly.
- **Verification after fix.** Build must pass. Invalid JSON-LD is
+- **Dispatcher verifies.** Build pass + invalid-JSON-LD revert happen in
-  reverted immediately.
+  the dispatcher after it applies the bundle — never in this agent.
 - **Transparency.** Every automated change logged in §14.
@@ -2,7 +2,7 @@
 name: plugin-advisor
 description: Plugin-fit checker — dispatched by /plugin-check and orchestrator gates (init-project, ship-feature). Recommends enable/disable.
 tools: Read, Bash, Glob, Grep
-model: haiku
+model: sonnet
 ---
 # PLUGIN ADVISOR
@@ -130,6 +130,9 @@ READY: <N> v1 features | entry points ✅ | config ✅ | CLAUDE.md ✅ | README
 ## PHASE 6 — DOC SYNC (automatic)
-Load `$HOME/.claude/agents/doc-syncer.md`.
+**INLINE-LOAD** `$HOME/.claude/agents/doc-syncer.md` — continue AS
-Execute in automatic mode:
+doc-syncer in THIS SAME context (you *become* it). This is an inline load,
 NOT a subagent dispatch: the `Agent` tool is not involved (which is why
 this agent correctly omits `Agent` from its `tools:`). Execute in
 automatic mode:
 `auto-mode scope: <list of all files created during scaffolding>`
@@ -2,6 +2,7 @@
 name: security-auditor
 description: SAST security gate — runs the pinned semgrep rulesets + the CLAUDE.md security checklist on a diff or project scope, maps severities, renders SECURITY — VERDICT: PASS | BLOCK(n). Blocks HIGH/CRITICAL only, reports the rest. Never fixes code. Fresh dispatch, no iteration history.
 tools: Read, Grep, Glob, Bash, Write
 model: sonnet
 ---
 # SECURITY-AUDITOR AGENT
@@ -1,7 +1,7 @@
 ---
 name: seo-analyzer
-description: Classical SEO audit agent (Google, Bing) — dispatched from /seo. Live audit: Core Web Vitals, on-page, technical, local SEO, legal (FR). Autonomous fixes + scored report. AI/GEO → geo-analyzer agent.
+description: Classical SEO audit agent (Google, Bing) — dispatched from /seo. Live audit: Core Web Vitals, on-page, technical, local SEO, legal (FR). Emits a fix bundle (dispatcher applies) + scored report. AI/GEO → geo-analyzer agent.
-tools: Read, Edit, Write, Bash, Grep, Glob, Agent, WebFetch, WebSearch
+tools: Read, Edit, Write, Bash, Grep, Glob, WebFetch, WebSearch
 ---
 # SEO — Classical Search Engines audit, fix & strategy
@@ -613,7 +613,7 @@ For each:
 - Description
 - Estimated time
 - Expected impact (high / medium / low)
- AUTO (executed in STEP 12) or USER (in SEO.md §11, with automation options)
+- AUTO (bundled in STEP 12, applied by the dispatcher) or USER (in SEO.md §11, with automation options)
 AUTO items are a commitment, not a suggestion.
@@ -689,80 +689,106 @@ Do not proceed to STEP 12 until this plan is printed.
 ---
-## STEP 12 — EXECUTE FIXES `[both]`
+## STEP 12 — EMIT FIX BUNDLE `[both]`
-**Orchestration step.** Delegate to specialist agents. Do NOT edit
+**You do NOT apply fixes and you do NOT dispatch any sub-agent.** Same
-files directly (except image pipeline).
+contract as `validator-analyzer`: you audit, then serialize the STEP 11
 batches into a machine-parseable FIX BUNDLE. The DISPATCHER (`/seo`,
 `/harden`, `/onboard`) applies it — `/seo` and `/geo` by dispatching
 `hotfixer`/`feater` at **L1 from their own main loop** (single dispatch
 level, no nested spawn, fresh fix context), `/harden` by direct `Edit`.
 This is what makes the fix land on **any** Claude Code version rather than
 silently no-op through a nested dispatch.
-### Batch A — Hotfixes (parallel when independent)
+Map every STEP 11 batch into the bundle tiers:
 | STEP 11 batch | Bundle tier | applier |
 |---|---|---|
 | A — Hotfixes | AUTO | hotfixer |
 | B — Small features | AUTO | feater |
 | C — Image pipeline | AUTO | bash |
 | D — Structural changes | GATED | feater |
 | E — Content removal | GATED | manual |
 | F — User actions | USER ACTIONS | — |
 ### Item requirements (self-contained)
 Every AUTO/GATED item MUST carry `id`, `applier`, `files`, and enough
 `current`/`expected` (or `change`/`impact`) detail for a **fresh**
 hotfixer/feater to act without re-auditing — it sees ONLY the item, never
 your audit context. Embed in each item:
 - **Shared-file edit discipline** — on shared templates (Layout.astro,
  index.html, base.html.twig…) instruct a narrow `Edit` on YOUR concern
  (meta tags) only; NEVER `Write`. `Write` only on sole-owned files
  (sitemap.xml, .htaccess, legal pages, new pages).
 - **Framework note** — Next.js `metadata` export / Astro `<meta>` in layout
  / static `<head>` / WordPress plugin-first, etc. (table below).
 - **Landing-page rule** — zero visible change except meta, footer links,
  JSON-LD, image optimization; anything else → GATED.
 - **Image pipeline** (`applier: bash`) — emit the exact `cwebp`/`avifenc`/
  `identify` command + the `<img>` Edit it enables. Do NOT run it yourself.
 ### Output shape
 ```
-Agent(subagent_type="hotfixer")
+## FIX BUNDLE (for dispatcher)
-prompt: "SEO hotfix: <fix description>.
+
-  File: <path>
+### AUTO — apply without confirmation
-  Current state: <what's wrong — specific lines>
+- id: A1
-  Expected state: <what it should be>
+  applier: hotfixer
-  Context: SEO audit fix, autonomous scope — no confirmation needed.
+  files: src/layouts/Base.astro
-  Do NOT commit — just fix and verify."
+  concern: <meta name="description"> missing
  current: <head> has no <meta name="description">
  expected: add <meta name="description" content="…"> (Astro — narrow Edit in layout <head>)
 - id: B1
  applier: feater
  files: src/pages/mentions-legales.astro, politique-confidentialite.astro, cgv.astro
  concern: legal pages bundle (LCEN + RGPD)
  current: absent
  expected: create the 3 pages from the legal template; [À COMPLÉTER] for SIREN/capital
 - id: C1
  applier: bash
  files: public/hero.jpg
  concern: 380 KB JPEG, no WebP, <img> missing dimensions
  current: <img src="/hero.jpg"> no width/height; hero.jpg 380KB
  expected: `cwebp -q 80 public/hero.jpg -o public/hero.webp`; then Edit <img> → add width/height from `identify -format "%wx%h"`
 ### GATED — apply only after user confirmation
 - id: D1
  applier: feater
  files: src/pages/ (new)
  change: 3 city landing pages (30/70 rule)
  impact: 3 new visible pages added to nav
 ### USER ACTIONS — never auto (report §11, each with automation-catalog ref)
 - Submit sitemap to Bing Webmaster Tools — automation: automation-catalog.md → IndexNow+Bing
 - GMB NAP correction — automation: <catalog ref>
 READY TO APPLY — awaiting dispatcher confirmation
 ```
-### Batch B — Small features (sequential)
+Emit the `READY TO APPLY — awaiting dispatcher confirmation` line **verbatim**
 as the last line of the bundle — the dispatcher keys its apply step on it.
 Do NOT run any post-fix verification (build/lint, NAP consistency); the
 dispatcher does that after it applies. Your job ends at the sentinel.
-Typical units (one `feater` call each):
+### Bundle completeness checklist (did every finding reach the bundle?)
 - **Legal pages bundle**: mentions-legales + politique-confidentialite + cgv
  (shared structure → one call)
 - **.htaccess bundle**: redirects + security headers (CSP, HSTS,
  X-Frame-Options, Referrer-Policy, X-Content-Type-Options) +
  custom 404 rule
 - **CMP install**: tarteaucitron.js integration across layouts
 - **Footer links**: legal/service/city links in footer component
 - **Sitemaps**: image sitemap + video sitemap if content exists
 - **i18n hreflang**: if multi-language, add reciprocal hreflang + x-default
-### Batch C — Image pipeline (direct Bash)
+- [ ] Meta/title/OG/canonical → AUTO (hotfixer)
-
+- [ ] JSON-LD LocalBusiness/Organization → AUTO (hotfixer/feater) — detailed GEO schema → geo-analyzer
-```bash
+- [ ] Image alt/dimensions → AUTO (hotfixer); compression → AUTO (bash) or §11 if tools absent
-# Check tools
+- [ ] robots.txt / sitemap.xml → AUTO (hotfixer) — AI-bot directives → geo-analyzer
-command -v cwebp &>/dev/null && echo "cwebp: available" || echo "cwebp: not found"
+- [ ] .htaccess security headers, image/video sitemap, hreflang → AUTO (feater)
-command -v avifenc &>/dev/null && echo "avifenc: available" || echo "avifenc: not found"
+- [ ] Legal pages, CMP, footer links → AUTO (feater)
-command -v identify &>/dev/null && echo "identify: available" || echo "identify: not found"
+- [ ] Heading hierarchy, noindex on technical pages → AUTO (hotfixer)
-
+- [ ] Unverifiable aggregateRating removal → AUTO (hotfixer); stock-photo testimonials → GATED (E)
-# Compression
+- [ ] Structural / new pages → GATED (D)
-# cwebp -q 80 <input> -o <output.webp>
+- [ ] Video transcripts, GMB, directories → USER ACTIONS (§11)
 # avifenc --min 0 --max 63 -s 0 <input> <output.avif>
 # Dimension extraction for missing width/height
 # identify -format "%wx%h" <image>  → edit the <img> tag
 ```
 If tools absent, document in SEO.md §11 as user action with automation
 catalog options.
 ### Batch D — Structural changes (confirmation gate)
 Present the batch D list:
 ```
 STRUCTURAL CHANGES — approval needed:
  D1. <description> — impact: <what changes visually>
  D2. ...
 Approve all / select specific / skip all?
 ```
 Approved → `feater` with detailed spec. Unapproved → SEO.md §9.
 ### Batch E — Content removal (confirmation gate)
 Same pattern as D.
 ### Batch F — User actions
 No execution. Documented in SEO.md §11 during STEP 13. Every entry
 MUST cite automation options from `~/.claude/agents/resources/automation-catalog.md`.
 ### Framework-specific notes
-Include in every sub-agent prompt:
+Carry the relevant note into each bundle item so the applier honors it:
 - **Next.js** — `metadata` export (App Router) or `Head` (Pages Router). `next-sitemap`. Redirects + headers in `next.config.js`.
 - **Astro** — direct `<meta>` in layouts. `@astrojs/sitemap`. Redirects in `astro.config.mjs` or `_redirects`.
@@ -790,48 +816,12 @@ Zero visible change on landing/homepage except:
 Anything else → batch D (confirmation).
-### Post-execution verification
+### Handoff to dispatcher
-1. **Syntax check** — HTML, JSON-LD, .htaccess
+Post-fix verification (build/lint, NAP consistency across JSON-LD /
-2. **Consistency check** — NAP matches across JSON-LD / visible / GMB
+visible / GMB, revert-on-break) and the §15 change log are the
-3. **No regressions**:
+DISPATCHER's responsibility, AFTER it applies the bundle at L1. You
-   ```bash
+emitted the bundle terminated by the sentinel — stop here.
   # npm run build, npm run lint, etc. — detect and run
   ```
 4. Broken sub-agent fix → revert.
 ### Execution checklist
 - [ ] Meta/title/OG/canonical → fixed (batch A)
 - [ ] JSON-LD LocalBusiness/Organization → fixed (batch A/B) — NOTE: detailed GEO schema audit handled by geo-analyzer
 - [ ] Image issues (alt, dimensions) → fixed (batch A)
 - [ ] Image compression → done/documented (batch C)
 - [ ] Video transcripts → documented (batch F, user action)
 - [ ] robots.txt / sitemap.xml → fixed (batch A) — AI-bot directives handled by geo-analyzer
 - [ ] Image/video sitemap → added if relevant (batch B)
 - [ ] .htaccess security headers → added (batch B)
 - [ ] Heading hierarchy → fixed (batch A)
 - [ ] hreflang if multi-language → fixed (batch A/B)
 - [ ] Legal pages → created (batch B)
 - [ ] CMP → installed (batch B)
 - [ ] noindex on technical pages → added (batch A)
 - [ ] Footer links → added (batch B)
 - [ ] Unverifiable aggregateRating → removed (batch A)
 - [ ] Stock photo testimonials → flagged (batch E)
 - [ ] Structural changes → approved items done (batch D)
 ### Change log
 ```
 BATCH: <A/B/C/D>
 AGENT: <hotfixer/feater/bash>
 FILE: <path>
 CHANGE: <what>
 REASON: <SEO rule or legal requirement>
 VERIFIED: <yes — how / no — why>
 ```
 All logs → SEO.md §15.
 ---
@@ -868,7 +858,11 @@ SEO AGENT RESULT (depth: <LOCAL|FULL>)
 ## ENTRIES FOR SEO.md §9 (medium term):
 ## ENTRIES FOR SEO.md §10 (long term):
 ## ENTRIES FOR SEO.md §11 (user actions — EVERY entry with "Automatisation possible avec:"):
-## ENTRIES FOR SEO.md §15 (change log):
+## ENTRIES FOR SEO.md §15 (change log — filled by the DISPATCHER after it applies the bundle):
 ## FIX BUNDLE (for dispatcher):
 <the AUTO / GATED / USER ACTIONS block from STEP 12, ending with the
 verbatim `READY TO APPLY — awaiting dispatcher confirmation` sentinel>
 ## SEO SCORING:
 <Scoring block from STEP 9>
@@ -938,26 +932,28 @@ PROCHAINE ETAPE : <highest-priority>
 ## RULES
 ### Orchestration
- **Analyze before fixing.** STEPs 0-11 pure analysis. No file
+- **Analyze, then bundle — never apply.** STEPs 0-11 are analysis;
-  modification until STEP 12.
+  STEP 12 emits a FIX BUNDLE. You NEVER edit a code file (report files
- **Delegate to specialists.** Never edit files directly in STEP 12
+  only) and NEVER dispatch a sub-agent. The dispatcher applies the
-  (except image pipeline). `hotfixer` for 1-2 file fixes, `feater`
+  bundle at L1 — this is the single-dispatch-level contract that makes
-  for multi-file features.
+  fixes land on any Claude Code version (no nested spawn).
 - **Bundle items are self-contained.** Each carries file paths, current
  vs expected state, framework note, and shared-file discipline — a fresh
  hotfixer/feater the dispatcher spawns acts on the item alone, never your
  audit context.
 - **Depth-aware.** LOCAL skips STEPs 3-7. Same rigor on what does run.
 - **Sub-agent prompts self-contained.** File paths, line numbers,
  current state, expected state, framework context, business context.
  Never assume sub-agent has audit findings.
 - **Do not audit GEO.** Detailed AI-crawler directives, llms.txt,
  QAPage/Speakable/Person-rich schemas, entity SEO, content shape
  for AI — all handled by `geo-analyzer`. Reference by name when needed.
 ### Scope
- **Autonomous fixes = markup, assets, config, legal pages.** Never
+- **Bundle-able scope = markup, assets, config, legal pages.** Never
  change business logic, layout, styles, routing unless confirmed.
 - **Shared-file edit discipline.** On template files shared with
  `geo-analyzer` (Layout.astro, index.html, base.html.twig, etc.),
-  your sub-agents (`hotfixer`/`feater`) MUST use `Edit` with a narrow
+  each bundle item MUST instruct the applier (`hotfixer`/`feater`) to
-  `old_string` targeting ONLY your owned concern (meta tags). NEVER
+  use `Edit` with a narrow `old_string` targeting ONLY your owned
  concern (meta tags). NEVER
  `Write` on shared templates. `Write` is reserved for files you
  solely own: sitemap.xml, .htaccess, legal pages, new city/service
  pages. Full-template refactor → escalate as user action in §11.
@@ -986,4 +982,5 @@ PROCHAINE ETAPE : <highest-priority>
 - **Iterative SEO.md.** Preserve Historique section.
 - **Transparency.** Every automated change logged with file, change,
  reason.
- **Verify after fix.** Build/lint must pass. Broken fixes reverted.
+- **Dispatcher verifies.** Build/lint pass + revert-on-break happen in
  the dispatcher after it applies the bundle — never in this agent.
@@ -2,6 +2,7 @@
 name: verifier
 description: Fresh independent verifier — reads a CONTRACT file from disk and renders a structured verdict (CONFORME / ECARTS / ERROR) on the implemented diff. Report-only, never fixes. Dispatched fresh at every iteration; receives no iteration history.
 tools: Read, Grep, Glob, Bash
 model: sonnet
 ---
 # VERIFIER AGENT
@@ -20,18 +20,85 @@ allowed-tools:
  - WebSearch
 ---
-Load and follow strictly:
+# /geo — GEO (AI-search) audit + fix dispatcher
 - $HOME/.claude/agents/geo-analyzer.md
-Execute the GEO-ANALYZER agent on the following target:
+Dispatches the `geo-analyzer` subagent (audit + fix bundle), then applies
 the bundle from THIS main loop at **L1** — same shape as `/web-validate`
 and `/seo`. The analyzer never edits files: it emits a `## FIX BUNDLE`
 terminated by `READY TO APPLY — awaiting dispatcher confirmation`, and this
 skill applies it. Applying from here (one dispatch level, no nested spawn)
 is what makes fixes land on any Claude Code version.
 ## STEP 1 — Dispatch geo-analyzer (audit + bundle)
 ```
 Agent(subagent_type="geo-analyzer")
 prompt: """
 Dispatched from /geo. Execute your full spec at
 ~/.claude/agents/geo-analyzer.md (STEP 0 onward — gather depth + business
 context as needed; if you must ask the user, ask and I relay).
 Produce your report:
 - If .claude/audits/SEO.md already exists → merge findings into its
  §7 — Optimisation GEO / IA.
 - Else write .claude/audits/GEO.md.
 Then emit the `## FIX BUNDLE` (STEP 13) terminated by the verbatim
 `READY TO APPLY — awaiting dispatcher confirmation` sentinel. Do NOT apply
 any fix and do NOT dispatch any sub-agent — /geo applies your bundle.
 $ARGUMENTS
 """
 ```
 ## STEP 2 — Apply the fix bundle (from THIS main loop, at L1)
 The analyzer returned a `## FIX BUNDLE`. Apply it by dispatching
 `hotfixer`/`feater` at **L1** (one dispatch level, no nested spawn).
 **Skip this step if intervention mode = conservative (audit-only)** — leave
 the bundle in the report as ready-to-apply.
 **Tier recognition (tolerant of the analyzer's batch labels).** Classify by
 intent, not header wording: **AUTO** = no-confirmation items (G1–G4/G6);
 **GATED** = items marked NEEDS CONFIRMATION / visible (G5); **USER ACTIONS**
 = G7.
 ### AUTO tier — no confirmation
 For each AUTO item, dispatch its `applier` at L1, passing the item verbatim:
 ```
 Agent(subagent_type="hotfixer")     # or "feater" per the item's applier
 prompt: "<paste the bundle item: files, concern, current, expected,
  framework note + shared-file discipline>.
  Context: GEO audit fix, autonomous scope — no confirmation needed.
  Do NOT commit — apply and self-verify only."
 ```
 ### GATED tier — confirmation required
 Present every GATED item (G5.x) in ONE gate:
 ```
 GEO — gated content-shape changes need approval (visible):
  G5.1 <change> — impact: <visible change>
 Approve all / select (ids) / skip all?
 ```
 Apply approved items via `feater` at L1. Unapproved → report §9 (medium
 term). NEVER apply a GATED item before explicit approval.
 ### After applying
 1. Build/lint if available (`npm run build`, `npm run lint`) — revert any
   applied fix that breaks the build; invalid JSON-LD reverted immediately.
 2. Record each applied change in the report change-log section.
 3. USER ACTIONS from the bundle → report §11 (each with automation-catalog ref).
 ## Note on integration
-If `.claude/audits/SEO.md` already exists, the geo-analyzer will
+If `.claude/audits/SEO.md` already exists, geo-analyzer merges its findings
-merge its findings into that file's `§7 — Optimisation GEO / IA`
+into that file's `§7 — Optimisation GEO / IA` section rather than writing a
-section (rather than writing a separate `GEO.md`). This keeps a
+separate `GEO.md`. This keeps a single consolidated report when both /seo
-single consolidated report when both /seo and /geo have been run.
+and /geo have been run.
 If no `.claude/audits/SEO.md` exists, the agent writes `.claude/audits/GEO.md` (run `mkdir -p .claude/audits` first).
@@ -135,22 +135,19 @@ typically contains BOTH concerns simultaneously:
  - meta tags (seo-analyzer)
  - JSON-LD blocks (geo-analyzer)
-When the agents' sub-agents (hotfixer/feater) run in parallel they
+The analyzers only AUDIT in parallel (read-only, safe). Fixes are applied
-could both target the same physical file. To avoid a `Write`-based
+LATER and SERIALLY by this dispatcher in STEP 1.5 (seo bundle first, then
-last-writer-wins scenario:
+geo bundle) — there is no parallel last-writer-wins race. Each bundle item
 still carries this rule for its applier:
-**Rule** (embedded in both agent dispatch prompts below):
+> On any shared template file (multiple owned concerns), use the `Edit`
 > tool with a **narrow, targeted** `old_string` enclosing ONLY the owned
 > concern. NEVER use `Write` (full-file rewrite) on a shared template.
 > `Write` is reserved for sole-owned files (sitemap.xml, robots.txt,
 > llms.txt, legal pages, new city pages, .htaccess).
-> On any shared template file (anything containing multiple owned
+If `Edit` is insufficient (full-template refactor), the item is escalated
-> concerns), use the `Edit` tool with a **narrow, targeted** `old_string`
+as a cross-agent note → §11 user action instead.
 > that encloses ONLY your owned concern. NEVER use `Write` (full-file
 > rewrite) on a shared template. `Write` is reserved for files you
 > are the sole owner of (sitemap.xml, robots.txt, llms.txt, legal
 > pages, new city pages, .htaccess).
 If a sub-agent determines `Edit` is insufficient (e.g. full template
 refactor needed), it must STOP and escalate as a cross-agent note —
 the dispatcher handles via §11 user action instead.
 ## STEP 1 — Spawn both agents IN PARALLEL
@@ -194,23 +191,23 @@ FILE OWNERSHIP (authoritative, prevents parallel-edit conflicts):
  Dispatcher escalates each note to SEO.md §11 as user action (with
  automation options). Do NOT attempt direct cross-agent fix.
-SHARED-FILE EDIT DISCIPLINE (last-writer-wins prevention):
+SHARED-FILE EDIT DISCIPLINE (carried into each bundle item):
 - On shared templates (Layout.astro, index.html, base.html.twig, etc.)
-  where meta tags + JSON-LD coexist, your sub-agents (hotfixer/feater)
+  where meta tags + JSON-LD coexist, each FIX BUNDLE item MUST instruct
-  MUST use `Edit` with a targeted `old_string` enclosing ONLY your
+  its applier (hotfixer/feater) to use `Edit` with a targeted `old_string`
-  concern (meta tags). NEVER use `Write` (full-file rewrite) on shared
+  enclosing ONLY your concern (meta tags). NEVER `Write` on shared templates.
-  templates.
+- `Write` is allowed only on sole-owned files: sitemap.xml, .htaccess,
- `Write` is allowed only on files where you are the sole owner:
+  legal pages, new city/service pages.
-  sitemap.xml, .htaccess, legal pages, new city/service pages.
+- If full-template refactor is needed, emit as a cross-agent note → §11.
 - If full-template refactor is needed, STOP and emit as a cross-agent
  note → user action in §11.
 Execute your agent spec at ~/.claude/agents/seo-analyzer.md starting
 at STEP 2 (skip STEP 0 and STEP 1 — context is provided above).
-At STEP 13, emit the STRUCTURED ENVELOPE for merging (not a
+At STEP 13, emit the STRUCTURED ENVELOPE for merging (not a standalone
-standalone SEO.md). Do NOT write any SEO.md file yourself — the
+SEO.md), INCLUDING the `## FIX BUNDLE` section terminated by the verbatim
-dispatcher will merge your output with geo-analyzer's output.
+`READY TO APPLY — awaiting dispatcher confirmation` sentinel. Do NOT apply
 any fix, do NOT dispatch any sub-agent, do NOT write SEO.md — /seo applies
 your bundle in STEP 1.5 and merges the reports.
 """
 Agent(subagent_type="geo-analyzer")
@@ -241,26 +238,86 @@ FILE OWNERSHIP (authoritative, prevents parallel-edit conflicts):
  Dispatcher escalates each note to SEO.md §11 as user action (with
  automation options). Do NOT attempt direct cross-agent fix.
-SHARED-FILE EDIT DISCIPLINE (last-writer-wins prevention):
+SHARED-FILE EDIT DISCIPLINE (carried into each bundle item):
 - On shared templates (Layout.astro, index.html, base.html.twig, etc.)
-  where meta tags + JSON-LD coexist, your sub-agents (hotfixer/feater)
+  where meta tags + JSON-LD coexist, each FIX BUNDLE item MUST instruct
-  MUST use `Edit` with a targeted `old_string` enclosing ONLY your
+  its applier (hotfixer/feater) to use `Edit` with a targeted `old_string`
-  concern (JSON-LD block). NEVER use `Write` (full-file rewrite) on
+  enclosing ONLY your concern (JSON-LD block). NEVER `Write` on shared
-  shared templates.
+  templates.
- `Write` is allowed only on files where you are the sole owner:
+- `Write` is allowed only on sole-owned files: robots.txt, llms.txt,
-  robots.txt, llms.txt, llms-full.txt.
+  llms-full.txt.
- If full-template refactor is needed, STOP and emit as a cross-agent
+- If full-template refactor is needed, emit as a cross-agent note → §11.
  note → user action in §11.
 Execute your agent spec at ~/.claude/agents/geo-analyzer.md starting
 at STEP 2 (skip STEP 0 and STEP 1 — context is provided above).
-At STEP 14, emit the STRUCTURED ENVELOPE for merging (not a
+At STEP 14, emit the STRUCTURED ENVELOPE for merging (not a standalone
-standalone GEO.md). Do NOT write any GEO.md or SEO.md file yourself —
+GEO.md), INCLUDING the `## FIX BUNDLE` section terminated by the verbatim
-the dispatcher will merge your output with seo-analyzer's output.
+`READY TO APPLY — awaiting dispatcher confirmation` sentinel. Do NOT apply
 any fix, do NOT dispatch any sub-agent, do NOT write GEO.md/SEO.md — /seo
 applies your bundle in STEP 1.5 and merges the reports.
 """
 ```
 ## STEP 1.5 — Apply fix bundles (from THIS main loop, at L1)
 Both analyzers returned an envelope containing a `## FIX BUNDLE` section
 terminated by `READY TO APPLY — awaiting dispatcher confirmation`. Apply
 them **from this dispatcher loop by dispatching `hotfixer`/`feater` at L1**
 — one dispatch level, no nested spawn, so fixes land on any Claude Code
 version (this is the whole point of the bundle contract).
 **Skip this step entirely if intervention mode = conservative (audit-only)**
 — leave both bundles in SEO.md as ready-to-apply and go to STEP 2.
 **Tier recognition (tolerant of the analyzer's batch labels).** Classify by
 intent, not header wording: **AUTO** = no-confirmation items (seo batches
 A/B/C · geo G1–G4/G6); **GATED** = items marked NEEDS CONFIRMATION / visible
 / structural (seo D/E · geo G5); **USER ACTIONS** = batch F / G7.
 ### Serial by ownership (no parallel race)
 The two bundles may touch the same shared template (meta vs JSON-LD). Apply
 **serially, never in parallel**:
 1. seo-analyzer AUTO items first (meta, sitemap, .htaccess, legal, images…).
 2. then geo-analyzer AUTO items (robots.txt, JSON-LD, llms.txt…).
 ### AUTO tier — no confirmation
 For each AUTO item, dispatch its `applier` at L1, passing the item verbatim:
 ```
 Agent(subagent_type="hotfixer")     # or "feater" per the item's applier
 prompt: "<paste the bundle item: files, concern, current, expected,
  framework note + shared-file discipline>.
  Context: SEO/GEO audit fix, autonomous scope — no confirmation needed.
  Do NOT commit — apply and self-verify only."
 ```
 `applier: bash` items → run the emitted command from this loop, then apply
 the `<img>` Edit it enables.
 ### GATED tier — confirmation required
 Collect every GATED item from BOTH bundles and present ONE gate:
 ```
 SEO/GEO — gated changes need approval (visible / structural):
  D1   <change> — impact: <visible change>            [seo]
  G5.1 <change> — impact: <visible change>            [geo]
 Approve all / select (ids) / skip all?
 ```
 Apply approved items via `feater` at L1 (same as AUTO). Unapproved →
 document in SEO.md §9. NEVER apply a GATED item before explicit approval.
 ### After applying
 1. Build/lint if available (`npm run build`, `npm run lint`) — revert any
   applied fix that breaks the build.
 2. Record each applied change for SEO.md §15 (file, change, reason, verified).
 3. USER ACTIONS from both bundles → SEO.md §11 (each with automation-catalog ref).
 ## STEP 2 — Merge envelopes into SEO.md
 Both agents return structured envelopes keyed by SEO.md section