fix(agent-loop): restore planning-intake gate admission

.agent-loop/initiatives/WS-ENG-ROOT-001-planning-intake-gate-recovery/STATUS.md

@@ -0,0 +1,9 @@
+# Status: WS-ENG-ROOT-001 — Planning-Intake Gate Recovery
+
+- Phase: one-use root recovery
+- Active planning chunk: none
+- Active implementation chunk: none
+- Recovery target: `WS-ENG-ROOT-001-01`
+- Signed basis and first parent: `339248c40020658583bf7bd1e4a58daf85f5ffb8`
+- Purpose: restore the already-documented closed first-planning-intake admission
+- Stop: recovery certificate is consumed by this exact merge and cannot replay

.agent-loop/initiatives/WS-ENG-ROOT-001-planning-intake-gate-recovery/chunks/WS-ENG-ROOT-001-01-exact-planning-intake-gate-recovery.md

@@ -0,0 +1,109 @@
+# Chunk Contract: WS-ENG-ROOT-001-01 — Exact Root Planning-Intake Gate Recovery
+
+## Goal
+
+Repair the circular trusted gates that reject the repository's documented
+first-new-initiative planning intake, using one exact consumed recovery.
+
+## Risk class
+
+L0
+
+## Start phase
+
+`implementation`
+
+## Machine-checkable scope
+
+```chunk-scope-json
+{
+  "schema_version": 1,
+  "chunk_id": "WS-ENG-ROOT-001-01",
+  "phase": "implementation",
+  "risk_class": "L0",
+  "allowed_paths": [
+    ".agent-loop/policies/loop-memory-recovery.json",
+    ".agent-loop/initiatives/WS-ENG-ROOT-001-planning-intake-gate-recovery/**",
+    ".agent-loop/merge-intents/WS-ENG-ROOT-001-01.json",
+    "scripts/check_chunk_contract.py",
+    "scripts/check_internal_review_evidence.py",
+    "scripts/check_loop_memory_state.py",
+    "scripts/update_post_merge_memory.py",
+    "scripts/test_agent_gates.py",
+    "scripts/test_check_chunk_contract.py",
+    "scripts/test_check_loop_memory_state.py",
+    "scripts/test_update_post_merge_memory.py"
+  ],
+  "forbidden_paths": ["backend/**", "frontend/**", ".github/**"],
+  "required_reviewers": ["senior engineering", "qa/test", "security/auth", "product/ops", "architecture", "ci integrity", "docs", "reuse/dedup", "test delta"],
+  "verification_commands": ["agent-gate-tests", "loop-memory-state", "loop-memory-recovery-tests", "chunk-scope-tests", "internal-review-evidence", "markdown-links", "stale-wording", "git-diff-check"]
+}
+```
+
+## Allowed files
+
+```text
+.agent-loop/policies/loop-memory-recovery.json
+.agent-loop/initiatives/WS-ENG-ROOT-001-planning-intake-gate-recovery/**
+.agent-loop/merge-intents/WS-ENG-ROOT-001-01.json
+scripts/check_chunk_contract.py
+scripts/check_internal_review_evidence.py
+scripts/check_loop_memory_state.py
+scripts/update_post_merge_memory.py
+scripts/test_agent_gates.py
+scripts/test_check_chunk_contract.py
+scripts/test_check_loop_memory_state.py
+scripts/test_update_post_merge_memory.py
+```
+
+## Not allowed
+
+```text
+product code, backend, frontend, workflows, dependencies, coverage weakening
+general unsigned implementation admission or reusable recovery authority
+automatic start, merge, approval, cancellation, or AUTH implementation
+```
+
+## Acceptance criteria
+
+- [ ] Only a brand-new initiative's additive canonical planning tree and one
+      `<initiative>-PLAN` intent can use the restored planning-intake admission.
+- [ ] Existing initiatives, implementation/configuration paths, scripts,
+      workflows, policies, deletes, renames, links, and executable modes fail.
+- [ ] Internal evidence accepts PLAN identity without inventing a PLAN contract,
+      while retaining exact reviewed-SHA and all required tracks.
+- [ ] Ordinary implementation/specification chunks remain signed-start-only.
+- [ ] Schema-v7 recovery is exact to this identity, signed basis, first parent,
+      empty recovered list, fixed reason/code, and is consumed on first use.
+- [ ] Independent state validation rejects altered recovery evidence.
+- [ ] Exactly one null-successor merge intent completes the root repair.
+
+## Verification commands
+
+```bash
+python3 scripts/test_check_chunk_contract.py
+python3 scripts/test_agent_gates.py
+PYTEST_DISABLE_PLUGIN_AUTOLOAD=1 python3 -m pytest -q scripts/test_update_post_merge_memory.py scripts/test_check_loop_memory_state.py
+python3 scripts/check_loop_memory_state.py
+python3 scripts/check_internal_review_evidence.py
+python3 scripts/check_markdown_links.py
+python3 scripts/check_stale_workstream_wording.py
+git diff --check origin/main...HEAD
+```
+
+## Required reviewers
+
+- [ ] senior engineering
+- [ ] QA/test
+- [ ] security/auth
+- [ ] product/ops
+- [ ] architecture
+- [ ] CI integrity
+- [ ] docs
+- [ ] reuse/dedup
+- [ ] test delta
+
+## Stop conditions
+
+Stop if the recovery is reusable, changes product code, or admits anything
+beyond the closed planning-intake tree.

.agent-loop/initiatives/WS-ENG-ROOT-001-planning-intake-gate-recovery/reviews/WS-ENG-ROOT-001-01-external-review-response.md

@@ -0,0 +1,20 @@
+# WS-ENG-ROOT-001-01 External Review Response
+
+Comments addressed: one CodeRabbit fail-closed finding. Independent planning
+state validation now rejects a non-string `intent_path` with a stable failure
+instead of raising, with a regression test for `null` input.
+
+Comments deferred: CodeRabbit's generic docstring-coverage warning is not a
+changed-code defect and adding unrelated docstrings would violate this exact
+recovery scope. The optional wider recovery runbook remains separate work.
+
+Human decisions needed: explicit approval of PR #205 remains required because
+trusted `main` contains the circular Agent Gate being repaired.
+
+Commands rerun: focused loop-memory tests, chunk-contract tests, agent-gate
+tests, independent state validation, internal evidence, Markdown links, stale
+wording, and diff integrity.
+
+Remaining risks: the trusted-main Agent Gate must fail on this recovery PR; it
+cannot consume candidate code without self-authorizing the repair. The exact
+schema-v7 recovery is base-pinned, identity-pinned, path-pinned, and one-use.

.agent-loop/initiatives/WS-ENG-ROOT-001-planning-intake-gate-recovery/reviews/WS-ENG-ROOT-001-01-internal-review-evidence.md

@@ -0,0 +1,53 @@
+# WS-ENG-ROOT-001-01 Internal Review Evidence
+
+Reviewed code SHA: `60db7ce2d682319b0c14ec0920b6adc75654bab9`
+
+Reviewed at: `2026-07-26T14:00:00Z`
+
+Reviewer run IDs: `cr_senior`, `cr_qa`, `cr_security`, `cr_product`,
+`cr_arch`, `cr_ci`, `cr_docs`, `cr_reuse`, `cr_testdelta`
+
+## Commands Run
+
+```bash
+python3 scripts/test_agent_gates.py
+python3 scripts/check_loop_memory_state.py
+PYTEST_DISABLE_PLUGIN_AUTOLOAD=1 python3 -m pytest -q scripts/test_update_post_merge_memory.py scripts/test_check_loop_memory_state.py
+python3 scripts/test_check_chunk_contract.py
+python3 scripts/check_internal_review_evidence.py
+python3 scripts/check_markdown_links.py
+python3 scripts/check_stale_workstream_wording.py
+git diff --check origin/main...HEAD
+```
+
+## Reviewer Results
+
+| Reviewer | Result | Blocking findings | Notes |
+|---|---|---|---|
+| senior engineering | PASS WITH LOW RISKS | none | The malformed-path repair is minimal and retains exact recovery boundaries. |
+| QA/test | PASS WITH LOW RISKS | none | The `null` intent-path regression proves a controlled fail-closed result. |
+| security/auth | PASS | none | Ordinary implementation remains signed-start-only; recovery remains exact and one-use. |
+| product/ops | PASS AFTER FIXES | none | The only procedural finding was this stale evidence, now rebound to the reviewed SHA. |
+| architecture | PASS AFTER FIXES | none | The external-response path is explicitly re-reviewed as part of the closed certificate. |
+| CI integrity | PASS WITH LOW RISKS | none | No workflow, threshold, exclusion, dependency, or coverage command was weakened. |
+| docs | PASS AFTER FIXES | none | The external response is recorded separately and this evidence is rebound to the reviewed SHA. |
+| reuse/dedup | PASS WITH LOW RISKS | none | Validator duplication is deliberate independent verification, not a forked helper. |
+| test delta | PASS WITH LOW RISKS | none | No tests were removed, skipped, deselected, or weakened. |
+
+## Findings Resolved
+
+Valid findings addressed: yes
+
+Open sub-agent sessions: none
+
+The repaired implementation also rejects a non-string planning-intake
+`intent_path` with a stable failure instead of raising. All nine reviewer
+tracks reviewed the exact implementation SHA above. Their only procedural
+finding was stale evidence after the external-review repair; this evidence-only
+commit resolves it without changing the reviewed implementation.
+
+## Remaining Gate
+
+GitHub checks, external review, and the explicit human merge checkpoint remain.
+The low-risk documentation suggestions are intentionally deferred because the
+one-use recovery certificate permits only its exact closed path set.

.agent-loop/initiatives/WS-ENG-ROOT-001-planning-intake-gate-recovery/reviews/WS-ENG-ROOT-001-01-pr-trust-bundle.md

@@ -0,0 +1,39 @@
+# PR Trust Bundle: WS-ENG-ROOT-001-01
+
+## Goal
+
+Restore the documented first-planning-intake path after PR #203 accidentally
+made it impossible to enter, using one exact consumed root recovery.
+
+## Scope and behavior
+
+Only trusted engineering gates, their tests, the independent memory checker,
+and the exact recovery certificate change. Product behavior is unchanged.
+Ordinary work remains signed-start-only.
+
+## Why root recovery is required
+
+Trusted base code rejects both the repair and every normal planning intake, so
+no ordinary signed contract can authorize the correction. The exact recovery
+is bound to current signed main and this single identity, then consumed.
+
+## Human review focus
+
+- Closed planning tree grammar and signed-history absence.
+- No candidate implementation self-authorization.
+- Exact recovery certificate, consumption, and replay inertness.
+- Continued signed-start requirement for ordinary chunks.
+
+## Human merge ownership
+
+The owner must explicitly approve this specific repair PR. A failing old scope
+check is the defect being repaired and must not be represented as passing.
+
+## External review response
+
+CodeRabbit identified one valid fail-closed issue: a non-string planning-intake
+`intent_path` could raise during independent state validation. The repair now
+returns a controlled validation failure and includes an adversarial regression.
+The Backend rerun passed after its first attempt hit a transient quay.io MinIO
+image-pull timeout. The trusted-main Agent Gate remains the expected circular
+failure that this exact recovery repairs.

.agent-loop/merge-intents/WS-ENG-ROOT-001-01.json

@@ -0,0 +1,9 @@
+{
+  "schema_version": 2,
+  "initiative_id": "WS-ENG-ROOT-001",
+  "chunk_id": "WS-ENG-ROOT-001-01",
+  "chunk_title": "Exact Root Planning-Intake Gate Recovery",
+  "next_chunk_id": null,
+  "next_chunk_title": null,
+  "next_requires_explicit_start": true
+}

.agent-loop/policies/loop-memory-recovery.json

@@ -1,22 +1,9 @@
 {
   "activation": {
-    "chunk_id": "WS-ENG-007-00R6",
-    "initiative_id": "WS-ENG-007"
+    "chunk_id": "WS-ENG-ROOT-001-01",
+    "initiative_id": "WS-ENG-ROOT-001"
   },
-  "signed_basis": "bba4ba5f171a4438b072740707a5cf8bde49d9af",
-  "recovered_merges": [
-    {
-      "chunk_id": "WS-ART-001-PLAN2",
-      "initiative_id": "WS-ART-001",
-      "merge_sha": "03a05eeb8f129e0d5f226cc5c058965f43590a81",
-      "pr_number": 197
-    },
-    {
-      "chunk_id": "WS-AUTH-001-11",
-      "initiative_id": "WS-AUTH-001",
-      "merge_sha": "f670b7058c71ad4d11a68c6e242e9fe501ae3aaf",
-      "pr_number": 201
-    }
-  ],
-  "schema_version": 6
+  "signed_basis": "339248c40020658583bf7bd1e4a58daf85f5ffb8",
+  "recovered_merges": [],
+  "schema_version": 7
 }