create-quiver 0.13.0 → 0.14.1

package/specs/quiver-v28-pixel-quiver-feedback-reconciliation/slices/slice-00-reconciliation-and-evidence-freeze/slice.json

@@ -0,0 +1,71 @@
+{
+  "slice_id": "slice-00-reconciliation-and-evidence-freeze",
+  "ticket": "QUIVER-28-00",
+  "type": "docs",
+  "title": "Reconciliation and evidence freeze",
+  "objective": "Freeze Pixel Quiver evidence, reconcile every supplied finding against v27, and produce the final implementation matrix before code changes.",
+  "description": "This mandatory documentary slice prevents duplicate fixes by classifying every QP, QIS, and framework improvement as resolved, partial, pending, regression, duplicate, or out of scope with evidence.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v28-reconciliation-evidence-freeze",
+    "branch_name": "feature/QUIVER-28-00-v28-reconciliation-evidence-freeze"
+  },
+  "files": [
+    "specs/quiver-v28-pixel-quiver-feedback-reconciliation/**",
+    "README_FOR_AI.md",
+    "ROADMAP.md",
+    "CHANGELOG.md"
+  ],
+  "expected_read_paths": [
+    "README_FOR_AI.md",
+    "ROADMAP.md",
+    "BACKLOG.md",
+    "CHANGELOG.md",
+    "specs/quiver-v27-reliability-ai-workflow-hardening/**",
+    "src/create-quiver/**",
+    "tests/**"
+  ],
+  "allowed_write_paths": [
+    "specs/quiver-v28-pixel-quiver-feedback-reconciliation/**",
+    "README_FOR_AI.md",
+    "ROADMAP.md",
+    "CHANGELOG.md"
+  ],
+  "depends_on": [],
+  "parallel_safe": "no",
+  "parallel_safe_reason": "slice-00 is the mandatory foundation and controls what later slices are allowed to fix.",
+  "must": [
+    "Freeze input evidence with paths, timestamps or hashes, and source notes.",
+    "Compare Pixel Quiver findings against v27 implementation, docs, and tests.",
+    "Produce the final coverage matrix with one final status per finding.",
+    "Update docs only if needed to reflect planned state.",
+    "Do not modify product code."
+  ],
+  "not_included": [
+    "CLI behavior changes.",
+    "New tests beyond documentation validation.",
+    "npm publishing.",
+    "PR creation."
+  ],
+  "acceptance": [
+    "Every supplied QP/QIS/framework improvement is classified with evidence.",
+    "Every pending or partial finding is mapped to a primary implementation slice.",
+    "Every already-resolved finding has a test, code, or doc reference.",
+    "No implementation slice starts before this matrix is complete.",
+    "No product code is modified."
+  ],
+  "tests": [
+    "git diff --check",
+    "find specs/quiver-v28-pixel-quiver-feedback-reconciliation -name \"slice.json\" -print -exec node -e \"JSON.parse(require('fs').readFileSync(process.argv[1], 'utf8'))\" {} \\;",
+    "npx create-quiver spec validate specs/quiver-v28-pixel-quiver-feedback-reconciliation"
+  ],
+  "validation_hints": [
+    "Do not copy unsanitized absolute paths into fixtures.",
+    "Treat v27 claims as hypotheses until verified against code and tests."
+  ],
+  "estimated_hours": 4,
+  "actual_hours": 4,
+  "completed_at": "2026-05-25T15:19:49Z",
+  "status": "completed"
+}

package/specs/quiver-v28-pixel-quiver-feedback-reconciliation/slices/slice-01-ai-run-state-approvals-and-clean-output/CLOSURE_BRIEF.md

@@ -0,0 +1,38 @@
+# CLOSURE BRIEF - slice-01: AI run state, approvals, and clean output
+
+## Summary
+
+Completed the AI run-state and approval-visibility hardening slice. Successful provider-backed planner commands now print clean extracted output instead of raw stdout/stderr, run-scoped approvals are shown separately from global planner approval files, and stale runs can be closed without deleting evidence.
+
+## Validation
+
+Passed:
+
+- `node --test tests/commands/ai-run-state.test.js tests/commands/ai-plan.test.js tests/commands/ai-export.test.js tests/commands/cli-contract.test.js`
+- `node --test tests/lib/ai-run-state.test.js tests/lib/approvals.test.js tests/lib/ai-export-state.test.js`
+
+## Relevant Changes
+
+- Updated `src/create-quiver/commands/ai.js`.
+- Updated `src/create-quiver/lib/ai/run-state.js`.
+- Updated `tests/commands/ai-run-state.test.js`.
+- Updated `tests/commands/ai-plan.test.js`.
+- Updated v28 coverage/status/evidence documents.
+
+## Pending Work
+
+- `slice-02`: enforce and repair structured `spec.slices[]` technical-plan contracts.
+- `slice-03`: reconcile multi-source active-slice state and stale `ai inspect` next commands.
+- `slice-04`: align spec validation, scope, and worktree reliability.
+- `slice-05`: add structured review-plan closure metadata and agent DX improvements.
+
+## Remaining Risks
+
+- `ai run close` archives by setting phase/status to `closed`; it does not delete or migrate any run artifacts.
+- Existing historical `.quiver/approvals/*` files may still be orphaned by design; the CLI now labels them instead of mutating them.
+- More advanced run replacement UX can be added later if dogfooding shows `close` plus visible open-run warnings are not enough.
+
+## Future Recommendations
+
+- Keep future planner commands using clean output on success and raw output only on failure diagnostics.
+- Keep approval reports grouped by run when adding new phases or approval types.

package/specs/quiver-v28-pixel-quiver-feedback-reconciliation/slices/slice-01-ai-run-state-approvals-and-clean-output/EXECUTION_BRIEF.md

@@ -0,0 +1,53 @@
+# EXECUTION BRIEF - slice-01: AI run state, approvals, and clean output
+
+## Context
+
+Pixel Quiver observed stale run state, approval history mixed with active work, and noisy provider output in AI commands.
+
+## Objective
+
+Make AI run and approval state clear, scoped to the active run when requested, and safe for humans and agents to consume.
+
+## Scope
+
+- AI command lifecycle and approvals.
+- Clean draft persistence.
+- Raw provider log storage and redaction.
+- Tests for run/approval states.
+
+## Acceptance Criteria
+
+- Active run and historical approvals are distinguishable.
+- Clean drafts do not include prompt echo or raw provider logs.
+- Raw logs are stored separately and redacted.
+- Existing run/approval data remains readable.
+
+## Technical Plan Summary
+
+Add or refine run-state and approval helpers, update AI command output, and add fixtures/tests around old and active lifecycle data.
+
+## Suggested Execution Steps
+
+1. Read `slice-00` final matrix for exact findings assigned here.
+2. Add failing tests for stale/historical approvals and noisy output.
+3. Update run/approval resolution.
+4. Update raw artifact persistence and output formatting.
+5. Run focused tests and update closure evidence.
+
+## Restrictions
+
+- Do not implement spec generation changes in this slice.
+- Do not change provider credentials handling.
+- Do not store secrets.
+
+## Risks
+
+- Existing `.quiver/approvals` layouts may vary across versions.
+
+## Completion Checklist
+
+- [ ] Run-state tests pass.
+- [ ] Approval grouping tests pass.
+- [ ] Clean draft/raw log tests pass.
+- [ ] Documentation evidence updated.
+

package/specs/quiver-v28-pixel-quiver-feedback-reconciliation/slices/slice-01-ai-run-state-approvals-and-clean-output/slice.json

@@ -0,0 +1,83 @@
+{
+  "slice_id": "slice-01-ai-run-state-approvals-and-clean-output",
+  "ticket": "QUIVER-28-01",
+  "type": "feature",
+  "title": "AI run state, approvals, and clean output",
+  "objective": "Harden AI run lifecycle, approval visibility, and provider output persistence so scopes and drafts do not get mixed.",
+  "description": "Separates active and historical runs, groups approvals by run, persists clean drafts and raw logs separately, and prevents stale approvals from appearing current.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v28-ai-run-state-approvals-clean-output",
+    "branch_name": "feature/QUIVER-28-01-v28-ai-run-state-approvals-clean-output"
+  },
+  "files": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/run-state.js",
+    "src/create-quiver/lib/ai/artifacts.js",
+    "src/create-quiver/lib/ai/export-state.js",
+    "src/create-quiver/lib/approvals.js",
+    "tests/commands/ai-run-state.test.js",
+    "tests/commands/ai-plan.test.js",
+    "tests/commands/ai-export.test.js",
+    "tests/lib/ai-run-state.test.js",
+    "tests/lib/approvals.test.js",
+    "tests/lib/ai-export-state.test.js",
+    "specs/quiver-v28-pixel-quiver-feedback-reconciliation/**"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/**",
+    "src/create-quiver/lib/approvals.js",
+    "tests/commands/ai-*.test.js",
+    "tests/lib/ai-*.test.js"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/run-state.js",
+    "src/create-quiver/lib/ai/artifacts.js",
+    "src/create-quiver/lib/ai/export-state.js",
+    "src/create-quiver/lib/approvals.js",
+    "tests/commands/ai-run-state.test.js",
+    "tests/commands/ai-plan.test.js",
+    "tests/commands/ai-export.test.js",
+    "tests/lib/ai-run-state.test.js",
+    "tests/lib/approvals.test.js",
+    "tests/lib/ai-export-state.test.js",
+    "specs/quiver-v28-pixel-quiver-feedback-reconciliation/**"
+  ],
+  "depends_on": [
+    "slice-00-reconciliation-and-evidence-freeze"
+  ],
+  "parallel_safe": "after_dependencies",
+  "parallel_safe_reason": "Can run after slice-00 if no overlap with slice-04 is introduced.",
+  "must": [
+    "Distinguish active, historical, stale, and orphaned approvals.",
+    "Keep clean drafts separate from raw provider logs.",
+    "Redact raw logs before persistence.",
+    "Avoid breaking existing run directories."
+  ],
+  "not_included": [
+    "Technical-plan schema repair.",
+    "Active-slice reconciliation.",
+    "PR creation."
+  ],
+  "acceptance": [
+    "AI commands do not mix current run output with historical approvals.",
+    "Raw provider logs are persisted separately from useful drafts.",
+    "Tests cover old, current, stale, and incomplete run states.",
+    "JSON export remains parseable without human log cleanup."
+  ],
+  "tests": [
+    "node --test tests/commands/ai-run-state.test.js tests/commands/ai-plan.test.js tests/commands/ai-export.test.js",
+    "node --test tests/lib/ai-run-state.test.js tests/lib/approvals.test.js tests/lib/ai-export-state.test.js"
+  ],
+  "validation_hints": [
+    "Add fixtures for old approvals before changing behavior.",
+    "Keep backward compatibility for existing .quiver/runs."
+  ],
+  "estimated_hours": 8,
+  "actual_hours": 8,
+  "completed_at": "2026-05-25T15:19:49Z",
+  "status": "completed"
+}

package/specs/quiver-v28-pixel-quiver-feedback-reconciliation/slices/slice-02-structured-technical-plan-contract-and-repair-flow/CLOSURE_BRIEF.md

@@ -0,0 +1,33 @@
+# CLOSURE BRIEF - slice-02: Structured technical plan contract and repair flow
+
+## Summary
+
+Completed the structured technical-plan contract and repair flow. Technical plans now explicitly prompt for `spec.slices[]`, approval blocks invalid technical-plan drafts before writing `approved.md`, and legacy approved plans missing structure can be repaired into a new draft with `ai repair-plan` while preserving the original approved artifact.
+
+## Validation
+
+Passed:
+
+- `node --test tests/commands/ai-plan.test.js tests/commands/ai-plan-spec-phase.test.js tests/commands/spec-create.test.js`
+- `node --test tests/commands/ai-review-plan.test.js tests/commands/flow.test.js tests/commands/cli-contract.test.js`
+- `node --test tests/lib/ai-spec-generator.test.js tests/lib/approvals.test.js`
+
+## Relevant Changes
+
+- `src/create-quiver/commands/ai.js` now validates technical-plan drafts before approval and implements `ai repair-plan`.
+- `src/create-quiver/lib/ai/spec-generator.js` exposes a shared technical-plan contract validator.
+- `src/create-quiver/index.js` registers and documents `ai repair-plan`.
+- `tests/commands/ai-plan.test.js` covers approval blocking, repair dry-run, repair execution, and approved artifact preservation.
+- Existing spec generation and flow tests were updated to use valid structured technical plans when approval is expected.
+
+## Pending Work
+
+None for this slice.
+
+## Remaining Risks
+
+- `ai repair-plan` depends on provider output quality. Invalid repair output is rejected and raw evidence is preserved, but no draft is written until the provider returns a valid structured contract.
+
+## Future Recommendations
+
+- Consider adding machine-readable repair metadata to `ai inspect` once slice-03 reconciles lifecycle state.

package/specs/quiver-v28-pixel-quiver-feedback-reconciliation/slices/slice-02-structured-technical-plan-contract-and-repair-flow/EXECUTION_BRIEF.md

@@ -0,0 +1,53 @@
+# EXECUTION BRIEF - slice-02: Structured technical plan contract and repair flow
+
+## Context
+
+Pixel Quiver reached an approved technical plan but `spec create` failed because the plan lacked a structured `spec.slices[]` array.
+
+## Objective
+
+Make the structured technical-plan contract explicit, validated, and repairable before spec creation.
+
+## Scope
+
+- Technical-plan prompts and validation.
+- Spec generation from approved plans.
+- Approval repair path.
+- Tests for valid, invalid, missing, and repaired structures.
+
+## Acceptance Criteria
+
+- Missing `spec.slices[]` is caught before writing spec files.
+- A valid plan creates the intended spec package.
+- A repair flow creates a derived draft without mutating the approved artifact.
+- Repaired drafts require review and approval.
+
+## Technical Plan Summary
+
+Add a formal minimum contract for `spec.slices[]`, validate it at approval/spec creation boundaries, and add a repair command or flow that preserves auditability.
+
+## Suggested Execution Steps
+
+1. Read final `slice-00` assignments.
+2. Add failing tests for missing `spec.slices[]`.
+3. Add schema/contract validation.
+4. Add repair flow.
+5. Validate dry-run no-write behavior.
+
+## Restrictions
+
+- Do not loosen approval gates.
+- Do not generate fallback generic slices silently.
+- Do not publish npm.
+
+## Risks
+
+- Planner output can vary; validation must produce helpful errors without requiring fragile formatting beyond the explicit contract.
+
+## Completion Checklist
+
+- [ ] Contract tests pass.
+- [ ] Repair tests pass.
+- [ ] Spec-create invalid input writes no files.
+- [ ] Closure evidence updated.
+

package/specs/quiver-v28-pixel-quiver-feedback-reconciliation/slices/slice-02-structured-technical-plan-contract-and-repair-flow/slice.json

@@ -0,0 +1,85 @@
+{
+  "slice_id": "slice-02-structured-technical-plan-contract-and-repair-flow",
+  "ticket": "QUIVER-28-02",
+  "type": "feature",
+  "title": "Structured technical plan contract and repair flow",
+  "objective": "Require a parseable technical-plan slice contract before spec creation and provide a safe repair path for approved plans with missing structure.",
+  "description": "Prevents spec creation from failing after human approval by validating `spec.slices[]` before approval/spec creation and by allowing a traceable derived repair draft.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v28-structured-plan-contract-repair-flow",
+    "branch_name": "feature/QUIVER-28-02-v28-structured-plan-contract-repair-flow"
+  },
+  "files": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/commands/spec.js",
+    "src/create-quiver/lib/ai/prompts.js",
+    "src/create-quiver/lib/ai/spec-generator.js",
+    "src/create-quiver/lib/ai/spec-templates.js",
+    "src/create-quiver/lib/approvals.js",
+    "tests/commands/ai-plan.test.js",
+    "tests/commands/ai-plan-spec-phase.test.js",
+    "tests/commands/spec-create.test.js",
+    "tests/lib/ai-spec-generator.test.js",
+    "tests/lib/approvals.test.js",
+    "specs/quiver-v28-pixel-quiver-feedback-reconciliation/**"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/lib/ai/**",
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/commands/spec.js",
+    "tests/commands/ai-*.test.js",
+    "tests/commands/spec-create.test.js",
+    "tests/lib/ai-spec-generator.test.js"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/commands/spec.js",
+    "src/create-quiver/lib/ai/prompts.js",
+    "src/create-quiver/lib/ai/spec-generator.js",
+    "src/create-quiver/lib/ai/spec-templates.js",
+    "src/create-quiver/lib/approvals.js",
+    "tests/commands/ai-plan.test.js",
+    "tests/commands/ai-plan-spec-phase.test.js",
+    "tests/commands/spec-create.test.js",
+    "tests/lib/ai-spec-generator.test.js",
+    "tests/lib/approvals.test.js",
+    "specs/quiver-v28-pixel-quiver-feedback-reconciliation/**"
+  ],
+  "depends_on": [
+    "slice-00-reconciliation-and-evidence-freeze",
+    "slice-01-ai-run-state-approvals-and-clean-output"
+  ],
+  "parallel_safe": "after_dependencies",
+  "parallel_safe_reason": "Can run after run/approval state is clarified.",
+  "must": [
+    "Define the minimum `spec.slices[]` contract.",
+    "Block approval or spec creation when required structure is missing.",
+    "Fail before writing files when the contract is invalid.",
+    "Provide a traceable repair path that creates a derived draft requiring review and approval."
+  ],
+  "not_included": [
+    "Active-slice reconciliation.",
+    "Worktree lifecycle changes.",
+    "Publishing npm."
+  ],
+  "acceptance": [
+    "A valid structured plan can create the expected spec/slices/handoffs.",
+    "An invalid approved plan fails before writing files with an actionable error.",
+    "Repairing an approved plan preserves the original and creates a derived draft.",
+    "Repair output requires review and approval before spec creation."
+  ],
+  "tests": [
+    "node --test tests/commands/ai-plan.test.js tests/commands/ai-plan-spec-phase.test.js tests/commands/spec-create.test.js",
+    "node --test tests/lib/ai-spec-generator.test.js tests/lib/approvals.test.js"
+  ],
+  "validation_hints": [
+    "Test both dry-run and real write modes.",
+    "Assert no files are written on invalid structure."
+  ],
+  "estimated_hours": 10,
+  "actual_hours": 10,
+  "status": "completed",
+  "completed_at": "2026-05-25T21:50:35Z"
+}

package/specs/quiver-v28-pixel-quiver-feedback-reconciliation/slices/slice-03-active-slice-reconciliation-and-ai-inspect/CLOSURE_BRIEF.md

@@ -0,0 +1,34 @@
+# CLOSURE BRIEF - slice-03: Active slice reconciliation and AI inspect
+
+## Summary
+
+Implemented active-slice reconciliation and spec-aware `ai inspect` recovery for the v28 Pixel Quiver feedback loop.
+
+## Validation
+
+Passed:
+
+- `node --test tests/commands/ai-export.test.js tests/commands/cli-contract.test.js`
+- `node --test tests/lib/project-state-resolver.test.js tests/lib/ai-export-state.test.js`
+
+## Relevant Changes
+
+- Added `collectActiveSliceState` in `src/create-quiver/lib/project-state-resolver.js` to read `docs/ai/ACTIVE_SLICE.md` and `ACTIVE_SLICES.md`.
+- Added reconciliation decisions: `preserve`, `close`, `replace`, and `blocked`.
+- Added `npx create-quiver ai active-slice status|reconcile`; `reconcile` is dry-run-first and currently refuses live writes.
+- Updated lifecycle export and inspect output to include active-slice state, blockers, dashboard data, and spec-aware next steps.
+- Added regression tests for conflicting active sources, missing active doc replacement, completed active slice closure, stale spec-create guidance, and CLI dry-run output.
+
+## Pending Work
+
+No pending work for this slice.
+
+## Remaining Risks
+
+- `ai active-slice reconcile` intentionally does not write changes yet. Future write mode should be implemented only with explicit safe file operations and backup behavior.
+- Active-slice Markdown parsing supports the current generated formats; unusual custom tables may still require manual correction.
+
+## Future Recommendations
+
+- Consider adding a future write-capable `ai active-slice reconcile --apply` only after the dry-run report is stable across real dogfooding projects.
+- Consider exposing active-slice state in any future dashboard/export consumer as a first-class readiness signal.

package/specs/quiver-v28-pixel-quiver-feedback-reconciliation/slices/slice-03-active-slice-reconciliation-and-ai-inspect/EXECUTION_BRIEF.md

@@ -0,0 +1,52 @@
+# EXECUTION BRIEF - slice-03: Active slice reconciliation and AI inspect
+
+## Context
+
+Pixel Quiver showed that active slice state can live in more than one file and that `ai inspect` can continue suggesting `spec create` after a spec exists by fallback.
+
+## Objective
+
+Make active-slice state inspectable, safely reconcilable, and reflected in AI inspection guidance.
+
+## Scope
+
+- Project state resolver.
+- AI inspect/export state.
+- Active slice status/reconciliation command surface if needed.
+- Tests and fixtures for conflicting active state.
+
+## Acceptance Criteria
+
+- All active-slice sources are detected.
+- Dry-run reconciliation is safe and explicit.
+- `ai inspect` does not recommend stale `spec create` when an existing spec should be validated or executed.
+- No worktrees or branches are deleted automatically.
+
+## Technical Plan Summary
+
+Extend the resolver to read active-slice sources, expose reconciliation states, and update AI inspect guidance.
+
+## Suggested Execution Steps
+
+1. Add fixtures for active-state conflicts.
+2. Extend resolver tests.
+3. Update inspect/export behavior.
+4. Add command/help updates if a new command surface is introduced.
+5. Validate no-write dry-run behavior.
+
+## Restrictions
+
+- Do not implement repair-plan logic.
+- Do not modify worktree lifecycle logic except for reading/reporting.
+
+## Risks
+
+- Existing generated projects may have partial active-slice files.
+
+## Completion Checklist
+
+- [ ] Resolver tests pass.
+- [ ] Inspect/export tests pass.
+- [ ] Dry-run behavior verified.
+- [ ] Docs/evidence updated.
+

package/specs/quiver-v28-pixel-quiver-feedback-reconciliation/slices/slice-03-active-slice-reconciliation-and-ai-inspect/slice.json

@@ -0,0 +1,82 @@
+{
+  "slice_id": "slice-03-active-slice-reconciliation-and-ai-inspect",
+  "ticket": "QUIVER-28-03",
+  "type": "feature",
+  "title": "Active slice reconciliation and AI inspect",
+  "objective": "Reconcile active slice state across multiple sources and make `ai inspect` recover from manual or fallback spec creation.",
+  "description": "Adds a safe, dry-run-first path for active-slice state and updates inspect guidance when specs already exist even if lifecycle state still points to spec creation.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v28-active-slice-reconciliation-ai-inspect",
+    "branch_name": "feature/QUIVER-28-03-v28-active-slice-reconciliation-ai-inspect"
+  },
+  "files": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/index.js",
+    "src/create-quiver/lib/project-state-resolver.js",
+    "src/create-quiver/lib/lifecycle.js",
+    "src/create-quiver/lib/statuses.js",
+    "src/create-quiver/lib/ai/export-state.js",
+    "tests/commands/ai-export.test.js",
+    "tests/commands/cli-contract.test.js",
+    "tests/lib/project-state-resolver.test.js",
+    "tests/lib/ai-export-state.test.js",
+    "specs/quiver-v28-pixel-quiver-feedback-reconciliation/**"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/lib/project-state-resolver.js",
+    "src/create-quiver/lib/lifecycle.js",
+    "src/create-quiver/commands/ai.js",
+    "tests/lib/project-state-resolver.test.js",
+    "tests/commands/ai-export.test.js"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/index.js",
+    "src/create-quiver/lib/project-state-resolver.js",
+    "src/create-quiver/lib/lifecycle.js",
+    "src/create-quiver/lib/statuses.js",
+    "src/create-quiver/lib/ai/export-state.js",
+    "tests/commands/ai-export.test.js",
+    "tests/commands/cli-contract.test.js",
+    "tests/lib/project-state-resolver.test.js",
+    "tests/lib/ai-export-state.test.js",
+    "specs/quiver-v28-pixel-quiver-feedback-reconciliation/**"
+  ],
+  "depends_on": [
+    "slice-00-reconciliation-and-evidence-freeze",
+    "slice-01-ai-run-state-approvals-and-clean-output"
+  ],
+  "parallel_safe": "after_dependencies",
+  "parallel_safe_reason": "Can run after the run/approval model is clarified.",
+  "must": [
+    "Detect active slice state in docs/ai/ACTIVE_SLICE.md and ACTIVE_SLICES.md.",
+    "Return safe states: preserve, close, replace, blocked.",
+    "Keep reconciliation dry-run-first.",
+    "Update ai inspect next steps when specs exist by fallback/manual creation."
+  ],
+  "not_included": [
+    "Spec-create structured plan contract.",
+    "Worktree filesystem validation.",
+    "Deleting branches or worktrees automatically."
+  ],
+  "acceptance": [
+    "Active-slice state is reported from every supported source.",
+    "Reconciliation dry-run lists planned file changes and risks.",
+    "No state is changed without explicit write mode.",
+    "ai inspect suggests spec validate/next/prompt-slice when a spec already exists."
+  ],
+  "tests": [
+    "node --test tests/commands/ai-export.test.js tests/commands/cli-contract.test.js",
+    "node --test tests/lib/project-state-resolver.test.js tests/lib/ai-export-state.test.js"
+  ],
+  "validation_hints": [
+    "Use fixtures for conflicting active slice files.",
+    "Do not delete real worktrees in tests."
+  ],
+  "estimated_hours": 10,
+  "actual_hours": 10,
+  "status": "completed",
+  "completed_at": "2026-05-25T22:02:27Z"
+}

package/specs/quiver-v28-pixel-quiver-feedback-reconciliation/slices/slice-04-spec-validation-scope-and-worktree-reliability/CLOSURE_BRIEF.md

@@ -0,0 +1,32 @@
+# CLOSURE BRIEF - slice-04: Spec validation, scope, and worktree reliability
+
+## Summary
+
+Completed the validation and worktree reliability hardening for v28. `spec validate` now catches missing execution git metadata, `spec status` reports expected but unregistered worktree paths as stale, `spec start --dry-run` gives actionable dirty-checkout recovery, and scope matching has regression coverage for exact paths plus supported globs.
+
+## Validation
+
+Passed:
+
+- `node --test tests/commands/spec-validate.test.js tests/commands/spec-worktree.test.js`
+- `node --test tests/lib/check-slice.test.js tests/lib/scope.test.js tests/lib/paths.test.js`
+
+## Relevant Changes
+
+- `src/create-quiver/commands/spec.js` validates required `git.*` execution metadata in every slice.
+- `src/create-quiver/lib/spec-worktrees.js` detects unregistered expected worktree directories and improves dirty-checkout recovery output.
+- `tests/commands/spec-validate.test.js` covers missing execution git metadata.
+- `tests/commands/spec-worktree.test.js` covers unregistered expected worktree paths and actionable dirty checkout output.
+- `tests/lib/scope.test.js` covers exact path plus supported glob write scopes.
+
+## Pending Work
+
+None for this slice.
+
+## Remaining Risks
+
+- Tightened validation may surface older specs that do not declare execution git metadata. This is intentional for local execution readiness and is covered by v28 release-readiness follow-up.
+
+## Future Recommendations
+
+- Consider adding a dedicated migration/check command for older specs if users report friction after release.