create-quiver 0.17.1 → 0.17.3

package/specs/quiver-v53-reliable-deep-project-analysis/slices/slice-04-safe-repair-layer/slice.json

@@ -0,0 +1,80 @@
+{
+  "slice_id": "slice-04-safe-repair-layer",
+  "ticket": "QUIVER-53-04",
+  "type": "reliability",
+  "title": "Safe repair layer",
+  "objective": "Repair safe provider schema drift without changing analysis semantics.",
+  "description": "Adds a minimal repair layer that removes unsupported additional properties such as notes, records every repair, and revalidates the analysis proposal before proceeding.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v53-safe-repair-layer",
+    "branch_name": "feature/QUIVER-53-04-v53-safe-repair-layer"
+  },
+  "files": [
+    "src/create-quiver/lib/ai/analyze-project-repair.js",
+    "src/create-quiver/lib/ai/analyze-project-validation.js",
+    "src/create-quiver/commands/ai.js",
+    "tests/lib/ai/analyze-project-repair.test.js",
+    "tests/commands/ai-analyze-project-provider.test.js",
+    "specs/quiver-v53-reliable-deep-project-analysis/STATUS.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/EVIDENCE_REPORT.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-04-safe-repair-layer/CLOSURE_BRIEF.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-04-safe-repair-layer/slice.json"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/lib/ai/analyze-project-schema.js",
+    "src/create-quiver/lib/ai/analyze-project-validation.js",
+    "tests/commands/ai-analyze-project-provider.test.js"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/lib/ai/analyze-project-repair.js",
+    "src/create-quiver/lib/ai/analyze-project-validation.js",
+    "src/create-quiver/commands/ai.js",
+    "tests/lib/ai/analyze-project-repair.test.js",
+    "tests/commands/ai-analyze-project-provider.test.js",
+    "tests/fixtures/analyze-project/**",
+    "specs/quiver-v53-reliable-deep-project-analysis/STATUS.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/EVIDENCE_REPORT.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-04-safe-repair-layer/CLOSURE_BRIEF.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-04-safe-repair-layer/slice.json"
+  ],
+  "depends_on": [
+    "slice-03-schema-error-grouping"
+  ],
+  "parallel_safe": "never",
+  "parallel_safe_reason": "Repair depends on validation classification and directly affects provider result flow.",
+  "must": [
+    "Repair only unsupported additionalProperties.",
+    "Remove extra keys such as notes without moving semantic content.",
+    "Record repair manifest entries with path, key, action, and reason.",
+    "Revalidate after repair.",
+    "Fail without final docs writes if repaired output remains invalid."
+  ],
+  "not_included": [
+    "Retry behavior.",
+    "Semantic rewriting.",
+    "Schema relaxation."
+  ],
+  "acceptance": [
+    "The notes drift fixture validates after safe repair.",
+    "Repair never changes confidence, evidence, claims, or business meaning.",
+    "Repair manifest records every removed key.",
+    "Unrepairable errors still fail without final docs writes."
+  ],
+  "tests": [
+    "node --test tests/lib/ai/analyze-project-repair.test.js",
+    "node --test tests/commands/ai-analyze-project-provider.test.js",
+    "npm run schema:slice:check",
+    "git diff --check"
+  ],
+  "validation_hints": [
+    "Assert that repair handles additionalProperties only."
+  ],
+  "estimated_hours": 6,
+  "actual_hours": 4,
+  "status": "completed",
+  "started_at": "2026-06-11T12:11:30Z",
+  "completed_at": "2026-06-11T12:16:13Z",
+  "blocked_reason": null
+}

package/specs/quiver-v53-reliable-deep-project-analysis/slices/slice-05-controlled-retry-layer/CLOSURE_BRIEF.md

@@ -0,0 +1,34 @@
+# CLOSURE_BRIEF - slice-05 controlled retry layer
+
+## Status
+
+Completed on 2026-06-11.
+
+## Summary
+
+Analyze-project now retries retryable invalid provider JSON with a default of 1 retry and a hard cap of 2 retries. Retry prompts use compact schema feedback plus the prior provider response rather than resending full selected file contents.
+
+## Evidence
+
+- Added retry prompt generation to `analyze-project-prompts`.
+- Added retry classification and retry loop to `runAnalyzeProject`.
+- Added retry manifests at `.quiver/runs/run-.../retry/analyze-project-retry.json`.
+- Provider tests cover retry success, retry exhaustion, and hard cap behavior.
+
+## Validation
+
+- `node --test tests/commands/ai-analyze-project-provider.test.js` passed: 17 tests.
+- `node --test tests/commands/ai-analyze-project.test.js` passed: 3 tests.
+- `node --test tests/lib/ai/analyze-project-repair.test.js` passed: 3 tests.
+- `npm run schema:slice:check` passed.
+- `git diff --check` passed.
+
+## Decisions
+
+- Provider command failures are treated as fatal and do not retry.
+- Retryable invalid JSON includes parse/schema/evidence/doc-path validation failures.
+- Retry prompts intentionally avoid `Selected file contents:` to keep retry payloads compact.
+
+## Follow-ups
+
+- `slice-06-audit-review-transaction` persists the final provider execution story.

package/specs/quiver-v53-reliable-deep-project-analysis/slices/slice-05-controlled-retry-layer/EXECUTION_BRIEF.md

@@ -0,0 +1,65 @@
+# EXECUTION_BRIEF - slice-05 controlled retry layer
+
+## Context
+
+Some provider outputs cannot be safely repaired but can be corrected by giving the provider compact validation feedback. Retry must be bounded to avoid cost, latency, and loops.
+
+## Objective
+
+Add controlled retry behavior for retryable analyze-project provider failures.
+
+## Scope
+
+- Retry classification.
+- Retry prompt.
+- Retry count/hard cap.
+- Retry output and manifest fields.
+- No final docs writes after exhausted retry.
+
+## Acceptance Criteria
+
+- Retry-success fixture succeeds.
+- Retry-exhausted fixture fails safely.
+- Fatal failures do not retry.
+- Retry count is visible and audited.
+
+## Expected Files To Modify
+
+- `src/create-quiver/commands/ai.js`
+- `src/create-quiver/lib/ai/analyze-project-prompts.js`
+- `src/create-quiver/lib/ai/analyze-project-validation.js`
+- `src/create-quiver/lib/ai/providers.js`
+- provider tests/fixtures
+
+## Validations Required
+
+- `node --test tests/commands/ai-analyze-project-provider.test.js`
+- `node --test tests/commands/ai-analyze-project.test.js`
+- `git diff --check`
+
+## Risks
+
+- Retry may increase command latency and provider cost.
+- Retrying fatal failures can hide root cause.
+
+## Dependencies
+
+- Depends on `slice-03-schema-error-grouping`.
+- Depends on `slice-04-safe-repair-layer`.
+
+## Instructions For Executor
+
+1. Keep retry count bounded and deterministic.
+2. Keep retry prompts compact.
+3. Do not write final docs until final JSON is valid.
+
+## Completion Checklist
+
+- [ ] Retry classification implemented.
+- [ ] Retry cap tests pass.
+- [ ] Fatal no-retry behavior is covered.
+- [ ] Closure brief updated.
+
+## Conditions Of Closure
+
+- Retry improves recovery without creating loops or unsafe writes.

package/specs/quiver-v53-reliable-deep-project-analysis/slices/slice-05-controlled-retry-layer/slice.json

@@ -0,0 +1,81 @@
+{
+  "slice_id": "slice-05-controlled-retry-layer",
+  "ticket": "QUIVER-53-05",
+  "type": "reliability",
+  "title": "Controlled retry layer",
+  "objective": "Retry provider analysis only when invalid output is retryable and bounded.",
+  "description": "Adds retry classification, compact retry prompts, retry count limits, timeout/error handling, and guarantees that exhausted retries do not write final docs.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v53-controlled-retry-layer",
+    "branch_name": "feature/QUIVER-53-05-v53-controlled-retry-layer"
+  },
+  "files": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/analyze-project-prompts.js",
+    "src/create-quiver/lib/ai/analyze-project-validation.js",
+    "src/create-quiver/lib/ai/providers.js",
+    "tests/commands/ai-analyze-project-provider.test.js",
+    "specs/quiver-v53-reliable-deep-project-analysis/STATUS.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/EVIDENCE_REPORT.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-05-controlled-retry-layer/CLOSURE_BRIEF.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-05-controlled-retry-layer/slice.json"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/analyze-project-prompts.js",
+    "src/create-quiver/lib/ai/providers.js",
+    "tests/commands/ai-analyze-project-provider.test.js"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/analyze-project-prompts.js",
+    "src/create-quiver/lib/ai/analyze-project-validation.js",
+    "src/create-quiver/lib/ai/providers.js",
+    "tests/commands/ai-analyze-project-provider.test.js",
+    "tests/fixtures/analyze-project/**",
+    "specs/quiver-v53-reliable-deep-project-analysis/STATUS.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/EVIDENCE_REPORT.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-05-controlled-retry-layer/CLOSURE_BRIEF.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-05-controlled-retry-layer/slice.json"
+  ],
+  "depends_on": [
+    "slice-03-schema-error-grouping",
+    "slice-04-safe-repair-layer"
+  ],
+  "parallel_safe": "never",
+  "parallel_safe_reason": "Retry changes provider execution flow and depends on repair outcomes.",
+  "must": [
+    "Retry only retryable errors.",
+    "Use default max retries of 1 and hard cap of 2 unless the contract changes.",
+    "Send compact schema feedback in retry prompts.",
+    "Avoid resending full context when not necessary.",
+    "Fail with 0 final docs writes after retry exhaustion."
+  ],
+  "not_included": [
+    "Unlimited retries.",
+    "Automatic fallback to another provider.",
+    "Cost accounting beyond available provider metadata."
+  ],
+  "acceptance": [
+    "Retry-success fixture succeeds after one retry.",
+    "Retry-exhausted fixture fails with compact output.",
+    "Fatal errors do not retry.",
+    "Retry count is visible in human output and manifests."
+  ],
+  "tests": [
+    "node --test tests/commands/ai-analyze-project-provider.test.js",
+    "node --test tests/commands/ai-analyze-project.test.js",
+    "git diff --check"
+  ],
+  "validation_hints": [
+    "Assert retry cap behavior exactly."
+  ],
+  "estimated_hours": 6,
+  "actual_hours": 4,
+  "status": "completed",
+  "started_at": "2026-06-11T12:16:20Z",
+  "completed_at": "2026-06-11T12:20:16Z",
+  "blocked_reason": null
+}

package/specs/quiver-v53-reliable-deep-project-analysis/slices/slice-06-audit-review-transaction/CLOSURE_BRIEF.md

@@ -0,0 +1,35 @@
+# CLOSURE_BRIEF - slice-06 audit and review transaction
+
+## Status
+
+Completed on 2026-06-11.
+
+## Summary
+
+Provider execution now leaves a redacted audit story under `.quiver/runs`, and review docs writes remain gated by editable review, final confirmation, snapshots, hashes, post-write validation, and atomic temp-file rename.
+
+## Evidence
+
+- Provider attempts persist redacted raw artifacts under `.quiver/runs/run-.../raw`.
+- Runs persist `status.json` with selected context, raw provider, repair, retry, and validation artifact references.
+- Review approval writes snapshots and hashes before final docs writes.
+- Final docs writes use temp files followed by rename.
+- Review cancel, no-TTY review, declined confirmation, and invalid edited JSON leave final docs unchanged while preserving audit artifacts.
+
+## Validation
+
+- `node --test tests/commands/ai-analyze-project-provider.test.js` passed: 17 tests.
+- `node --test tests/commands/ai-analyze-project-review.test.js` passed: 5 tests.
+- `node --test tests/commands/ai-analyze-project.test.js` passed: 3 tests.
+- `npm run smoke:create-quiver` passed.
+- `git diff --check` passed.
+
+## Decisions
+
+- Audit artifacts are allowed for real provider execution even when review is canceled; final docs remain unchanged.
+- Provider command failures do not retry, but still persist raw provider output and run status.
+- Atomic writes are implemented with a temporary sibling file and `renameSync`.
+
+## Follow-ups
+
+- `slice-07-semantic-validation-docs` will document and benchmark the completed reliability kernel.

package/specs/quiver-v53-reliable-deep-project-analysis/slices/slice-06-audit-review-transaction/EXECUTION_BRIEF.md

@@ -0,0 +1,71 @@
+# EXECUTION_BRIEF - slice-06 audit and review transaction
+
+## Context
+
+After provider execution, users need an audit trail and a safe review/write transaction. Provider-only runs may write redacted audit artifacts, but final docs must require valid JSON and review approval.
+
+## Objective
+
+Implement redacted audit artifacts and review/write transaction guarantees.
+
+## Scope
+
+- `.quiver/runs` artifacts.
+- Versioned manifests.
+- Edited proposal revalidation.
+- Final diff.
+- Snapshot and hash before write.
+- Atomic final docs writes.
+
+## Acceptance Criteria
+
+- Provider execution writes redacted audit artifacts.
+- `--dry-run` writes nothing.
+- Review cancel leaves docs unchanged.
+- Invalid edited JSON blocks writes.
+- Approved review writes snapshots and docs atomically.
+
+## Expected Files To Modify
+
+- `src/create-quiver/commands/ai.js`
+- `src/create-quiver/lib/ai/analyze-project-review.js`
+- `src/create-quiver/lib/ai/artifacts.js`
+- `src/create-quiver/lib/ai/run-state.js`
+- analyze-project provider/review tests
+
+## Validations Required
+
+- `node --test tests/commands/ai-analyze-project-provider.test.js`
+- `node --test tests/commands/ai-analyze-project-review.test.js`
+- `node --test tests/commands/ai-analyze-project.test.js`
+- `npm run smoke:create-quiver`
+- `git diff --check`
+
+## Risks
+
+- Audit files can leak sensitive content if redaction is incomplete.
+- Review flow can be confusing if audit writes are not distinguished from final docs writes.
+- Partial writes can leave docs inconsistent.
+
+## Dependencies
+
+- Depends on `slice-02-safe-context-boundary`.
+- Depends on `slice-04-safe-repair-layer`.
+- Depends on `slice-05-controlled-retry-layer`.
+
+## Instructions For Executor
+
+1. Use temporary files and atomic rename for final docs writes.
+2. Record hashes and snapshots before writing.
+3. Keep non-TTY behavior non-interactive and actionable.
+
+## Completion Checklist
+
+- [ ] Audit artifacts implemented.
+- [ ] Review cancel/approve tests pass.
+- [ ] Atomic write behavior is validated.
+- [ ] Closure brief updated.
+
+## Conditions Of Closure
+
+- Provider analysis has a reliable audit trail and docs writes are transactional.

package/specs/quiver-v53-reliable-deep-project-analysis/slices/slice-06-audit-review-transaction/slice.json

@@ -0,0 +1,88 @@
+{
+  "slice_id": "slice-06-audit-review-transaction",
+  "ticket": "QUIVER-53-06",
+  "type": "workflow",
+  "title": "Audit and review transaction",
+  "objective": "Persist provider execution evidence and make docs writes reviewable, atomic, and recoverable.",
+  "description": "Adds redacted run artifacts, validation/repair/retry manifests, review proposal handling, edited JSON revalidation, final diff, confirmation, snapshots, hashes, and atomic final docs writes.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v53-audit-review-transaction",
+    "branch_name": "feature/QUIVER-53-06-v53-audit-review-transaction"
+  },
+  "files": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/analyze-project-review.js",
+    "src/create-quiver/lib/ai/artifacts.js",
+    "src/create-quiver/lib/ai/run-state.js",
+    "tests/commands/ai-analyze-project-provider.test.js",
+    "tests/commands/ai-analyze-project-review.test.js",
+    "specs/quiver-v53-reliable-deep-project-analysis/STATUS.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/EVIDENCE_REPORT.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-06-audit-review-transaction/CLOSURE_BRIEF.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-06-audit-review-transaction/slice.json"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/analyze-project-review.js",
+    "src/create-quiver/lib/ai/artifacts.js",
+    "src/create-quiver/lib/ai/run-state.js",
+    "tests/commands/ai-analyze-project-review.test.js"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/analyze-project-review.js",
+    "src/create-quiver/lib/ai/artifacts.js",
+    "src/create-quiver/lib/ai/run-state.js",
+    "tests/commands/ai-analyze-project-provider.test.js",
+    "tests/commands/ai-analyze-project-review.test.js",
+    "tests/fixtures/analyze-project/**",
+    "specs/quiver-v53-reliable-deep-project-analysis/STATUS.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/EVIDENCE_REPORT.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-06-audit-review-transaction/CLOSURE_BRIEF.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-06-audit-review-transaction/slice.json"
+  ],
+  "depends_on": [
+    "slice-02-safe-context-boundary",
+    "slice-04-safe-repair-layer",
+    "slice-05-controlled-retry-layer"
+  ],
+  "parallel_safe": "never",
+  "parallel_safe_reason": "Audit and review transaction integrates safety, repair, retry, and final write behavior.",
+  "must": [
+    "Persist redacted raw output, repaired output when present, selected context, validation manifest, repair manifest, retry manifest, and final status.",
+    "Use versioned manifests with size bounds and truncation markers.",
+    "Ensure --review revalidates edited JSON before final diff.",
+    "Create snapshots and hashes before docs writes.",
+    "Write final docs atomically only after explicit approval."
+  ],
+  "not_included": [
+    "Implementing a restore command.",
+    "Writing docs without review approval.",
+    "Persisting unredacted provider secrets."
+  ],
+  "acceptance": [
+    "Provider execution creates redacted audit artifacts under .quiver/runs.",
+    "--dry-run creates no run artifacts.",
+    "Review cancel leaves final docs unchanged.",
+    "Approved review writes snapshots and final docs atomically.",
+    "Invalid edited JSON blocks writes."
+  ],
+  "tests": [
+    "node --test tests/commands/ai-analyze-project-provider.test.js",
+    "node --test tests/commands/ai-analyze-project-review.test.js",
+    "node --test tests/commands/ai-analyze-project.test.js",
+    "npm run smoke:create-quiver",
+    "git diff --check"
+  ],
+  "validation_hints": [
+    "Assert exact write counts for dry-run, provider-only, review-cancel, and review-approve flows."
+  ],
+  "estimated_hours": 8,
+  "actual_hours": 4,
+  "status": "completed",
+  "started_at": "2026-06-11T12:20:20Z",
+  "completed_at": "2026-06-11T12:23:37Z",
+  "blocked_reason": null
+}

package/specs/quiver-v53-reliable-deep-project-analysis/slices/slice-07-semantic-validation-docs/CLOSURE_BRIEF.md

@@ -0,0 +1,32 @@
+# CLOSURE_BRIEF - slice-07 semantic validation, docs, and benchmark
+
+## Status
+
+Completed on 2026-06-11.
+
+## Summary
+
+The evidence-quality validation already enforced by the parser is now documented alongside repair, retry, audit, review, and benchmark expectations for reliable `ai analyze-project`.
+
+## Evidence
+
+- `docs/reference/commands.md` documents dry-run, provider, repair, retry, audit, review, and benchmark behavior.
+- `docs/workflows/existing-project.md` adds direct steps for deep AI project analysis in existing projects.
+- `docs/TROUBLESHOOTING.md` explains provider JSON/schema failures and safe next steps.
+- Provider fixtures remain the deterministic benchmark path; live `nika-erp` smoke is documented as optional release evidence.
+
+## Validation
+
+- `node --test tests/commands/ai-analyze-project-provider.test.js` passed: 17 tests.
+- `npm run docs:check` passed.
+- `npm run smoke:create-quiver` passed.
+- `git diff --check` passed.
+
+## Decisions
+
+- No new semantic validator was added because existing parser validation already rejects evidence-poor non-unknown claims and downgrades confirmed claims backed by truncated evidence.
+- Live provider benchmark remains optional, not a CI gate.
+
+## Follow-ups
+
+- `slice-08-structural-map-hardening` can improve analysis quality after reliability is stable.

package/specs/quiver-v53-reliable-deep-project-analysis/slices/slice-07-semantic-validation-docs/EXECUTION_BRIEF.md

@@ -0,0 +1,66 @@
+# EXECUTION_BRIEF - slice-07 semantic validation, docs, and benchmark
+
+## Context
+
+Passing JSON schema is necessary but not sufficient. The analysis must remain evidence-backed and understandable to users.
+
+## Objective
+
+Add semantic validation and document the reliable analyze-project workflow.
+
+## Scope
+
+- Confidence/evidence validation.
+- Strict-mode behavior.
+- Command docs.
+- Troubleshooting docs.
+- Benchmark guidance.
+
+## Acceptance Criteria
+
+- `confirmed` claims require explicit evidence.
+- Evidence-poor output is reported as warning or strict error.
+- Docs explain dry-run, provider, repair, retry, audit, and review behavior.
+- Benchmark guidance separates deterministic fixtures from optional live provider smoke.
+
+## Expected Files To Modify
+
+- `src/create-quiver/lib/ai/analyze-project-validation.js`
+- i18n messages
+- `docs/reference/commands.md`
+- `docs/workflows/existing-project.md`
+- `docs/TROUBLESHOOTING.md`
+- provider tests/fixtures
+
+## Validations Required
+
+- `node --test tests/commands/ai-analyze-project-provider.test.js`
+- `npm run docs:check`
+- `npm run smoke:create-quiver`
+- `git diff --check`
+
+## Risks
+
+- Over-strict semantic validation can reject useful inferred analysis.
+- Docs can drift from actual CLI behavior.
+
+## Dependencies
+
+- Depends on `slice-06-audit-review-transaction`.
+
+## Instructions For Executor
+
+1. Keep semantic validation conservative.
+2. Do not promote inferred claims to confirmed.
+3. Keep docs aligned with CLI help and tests.
+
+## Completion Checklist
+
+- [ ] Semantic validation implemented.
+- [ ] Docs updated.
+- [ ] Benchmark guidance added.
+- [ ] Closure brief updated.
+
+## Conditions Of Closure
+
+- Users understand and can verify reliable analyze-project behavior.

package/specs/quiver-v53-reliable-deep-project-analysis/slices/slice-07-semantic-validation-docs/slice.json

@@ -0,0 +1,86 @@
+{
+  "slice_id": "slice-07-semantic-validation-docs",
+  "ticket": "QUIVER-53-07",
+  "type": "validation-docs",
+  "title": "Semantic validation, docs, and benchmark",
+  "objective": "Ensure schema-valid analysis is also evidence-backed and documented for users.",
+  "description": "Adds semantic validation for confidence/evidence quality, updates command documentation, and documents deterministic/live benchmark expectations for reliable analyze-project behavior.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v53-semantic-validation-docs",
+    "branch_name": "feature/QUIVER-53-07-v53-semantic-validation-docs"
+  },
+  "files": [
+    "src/create-quiver/lib/ai/analyze-project-validation.js",
+    "src/create-quiver/lib/i18n/messages/en.js",
+    "src/create-quiver/lib/i18n/messages/es.js",
+    "docs/reference/commands.md",
+    "docs/workflows/existing-project.md",
+    "docs/TROUBLESHOOTING.md",
+    "tests/commands/ai-analyze-project-provider.test.js",
+    "specs/quiver-v53-reliable-deep-project-analysis/STATUS.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/EVIDENCE_REPORT.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-07-semantic-validation-docs/CLOSURE_BRIEF.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-07-semantic-validation-docs/slice.json"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/lib/ai/analyze-project-validation.js",
+    "docs/reference/commands.md",
+    "docs/workflows/existing-project.md",
+    "docs/TROUBLESHOOTING.md",
+    "tests/commands/ai-analyze-project-provider.test.js"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/lib/ai/analyze-project-validation.js",
+    "src/create-quiver/lib/i18n/messages/en.js",
+    "src/create-quiver/lib/i18n/messages/es.js",
+    "docs/reference/commands.md",
+    "docs/workflows/existing-project.md",
+    "docs/TROUBLESHOOTING.md",
+    "tests/commands/ai-analyze-project-provider.test.js",
+    "tests/fixtures/analyze-project/**",
+    "specs/quiver-v53-reliable-deep-project-analysis/STATUS.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/EVIDENCE_REPORT.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-07-semantic-validation-docs/CLOSURE_BRIEF.md",
+    "specs/quiver-v53-reliable-deep-project-analysis/slices/slice-07-semantic-validation-docs/slice.json"
+  ],
+  "depends_on": [
+    "slice-06-audit-review-transaction"
+  ],
+  "parallel_safe": "after_dependencies",
+  "parallel_safe_reason": "Requires the final reliability kernel behavior to document accurately.",
+  "must": [
+    "Validate that meaningful claims have confidence and evidence.",
+    "Prevent confirmed claims without explicit evidence.",
+    "Warn or error on semantic validation issues according to strict mode.",
+    "Document repair, retry, audit, and review behavior.",
+    "Document deterministic benchmark and optional live nika-erp smoke."
+  ],
+  "not_included": [
+    "Structural symbol map implementation.",
+    "Guaranteeing perfect domain inference.",
+    "Making live provider benchmark mandatory in CI."
+  ],
+  "acceptance": [
+    "Schema-valid but evidence-poor fixture produces semantic warnings or strict errors.",
+    "Docs explain mode behavior and safe next steps.",
+    "Benchmark expectations are documented and reproducible with provider fake.",
+    "Command reference remains synchronized."
+  ],
+  "tests": [
+    "node --test tests/commands/ai-analyze-project-provider.test.js",
+    "npm run docs:check",
+    "npm run smoke:create-quiver",
+    "git diff --check"
+  ],
+  "validation_hints": [
+    "Keep live provider benchmark optional and documented as release evidence."
+  ],
+  "estimated_hours": 6,
+  "actual_hours": 3,
+  "status": "completed",
+  "started_at": "2026-06-11T12:23:45Z",
+  "completed_at": "2026-06-11T12:26:24Z",
+  "blocked_reason": null
+}