create-quiver 0.15.1 → 0.15.3

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-02-approve-interactive-selection/EXECUTION_BRIEF.md

@@ -0,0 +1,53 @@
+# EXECUTION_BRIEF - slice-02 Interactive approval selection
+
+## Context
+
+Users currently need to run `ai approvals`, inspect versions manually, then pass `--version <n>`. The requested UX should behave like `ai agent set`: guide users in TTY and stay explicit in automation.
+
+## Objective
+
+Add TTY approval selection for `ai approve` when `--version` is omitted.
+
+## Scope
+
+- `src/create-quiver/commands/ai.js`
+- selector helpers if needed
+- approval tests
+
+## Acceptance Criteria
+
+- TTY acceptance approval without `--version` shows candidates.
+- TTY technical-plan approval without `--version` shows review-aware candidates.
+- No-TTY/CI requires `--version`.
+- Explicit `--version` behavior remains stable.
+- Cancellation writes no files.
+- Blocked candidates cannot be approved.
+
+## Technical Plan Summary
+
+Resolve candidates from slice-01, use `selectOption` for prompt-capable environments, and pass the selected version into the existing approval path.
+
+## Suggested Steps
+
+1. Add approval selection resolver.
+2. Wire resolver into `runApprove`.
+3. Add no-TTY guardrail message.
+4. Add tests for acceptance and technical-plan.
+5. Verify explicit version path remains unchanged.
+
+## Restrictions
+
+- Do not add `--review` support.
+- Do not weaken technical-plan validation.
+- Do not prompt in CI/no-TTY/JSON.
+
+## Risks
+
+- Prompt availability detection must match existing `ai agent set` behavior closely enough to avoid CI hangs.
+
+## Completion Checklist
+
+- [ ] TTY selector implemented.
+- [ ] No-TTY guardrail tested.
+- [ ] Cancellation tested.
+- [ ] Explicit `--version` tested.

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-02-approve-interactive-selection/slice.json

@@ -0,0 +1,79 @@
+{
+  "slice_id": "slice-02-approve-interactive-selection",
+  "ticket": "QUIVER-33-02",
+  "type": "feature",
+  "title": "Interactive approval selection",
+  "objective": "Let `ai approve --phase acceptance|technical-plan` guide TTY users when `--version` is omitted while preserving explicit non-interactive behavior.",
+  "description": "Adds a selector patterned after `ai agent set`, recommends the current latest draft, blocks unsafe approvals, and keeps CI/no-TTY requiring `--version <n>`.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v33-approve-interactive-selection",
+    "branch_name": "feature/QUIVER-33-02-v33-approve-interactive-selection"
+  },
+  "files": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/cli/selectors.js",
+    "src/create-quiver/lib/approvals.js",
+    "tests/commands/ai-review-plan.test.js",
+    "tests/commands/ai-plan.test.js",
+    "tests/lib/cli-selectors.test.js"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/approvals.js",
+    "src/create-quiver/lib/cli/selectors.js",
+    "tests/commands/ai-review-plan.test.js",
+    "tests/commands/ai-plan.test.js",
+    "SPEC.md"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/cli/selectors.js",
+    "src/create-quiver/lib/approvals.js",
+    "tests/commands/ai-review-plan.test.js",
+    "tests/commands/ai-plan.test.js",
+    "tests/lib/cli-selectors.test.js",
+    "specs/quiver-v33-approval-ux-and-planner-progress/**"
+  ],
+  "depends_on": [
+    "slice-01-approval-candidates-model"
+  ],
+  "parallel_safe": "no",
+  "parallel_safe_reason": "This slice changes approval command behavior and should land before downstream consumers rely on it.",
+  "must": [
+    "Open a selector only when `--version` is omitted and prompts are available.",
+    "Require explicit `--version <n>` in CI/no-TTY.",
+    "Keep explicit `--version <n>` behavior backward compatible.",
+    "Recommend the current/latest approvable draft.",
+    "Show concise useful decision information for each candidate.",
+    "Allow cancellation without writes.",
+    "Do not approve older or blocked drafts silently."
+  ],
+  "not_included": [
+    "Changing plan-review result derivation.",
+    "Changing spec generation.",
+    "Adding editor review to approve."
+  ],
+  "acceptance": [
+    "`ai approve --phase acceptance` in TTY without `--version` shows a selector.",
+    "`ai approve --phase technical-plan` in TTY without `--version` shows a selector with review context.",
+    "No-TTY/CI without `--version` fails actionably.",
+    "`--version 1` still approves through the existing path when valid.",
+    "Canceling the selector writes no approval.",
+    "Older draft versions cannot be approved through the selector."
+  ],
+  "tests": [
+    "node --test tests/commands/ai-plan.test.js tests/commands/ai-review-plan.test.js tests/lib/cli-selectors.test.js",
+    "git diff --check"
+  ],
+  "validation_hints": [
+    "Use injected selector helpers in tests; do not require a real terminal.",
+    "Test no-TTY and CI explicitly."
+  ],
+  "estimated_hours": 8,
+  "status": "completed",
+  "blocked_reason": null,
+  "actual_hours": 8,
+  "completed_at": "2026-05-28"
+}

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-03-technical-plan-review-decision-data/CLOSURE_BRIEF.md

@@ -0,0 +1,19 @@
+# CLOSURE_BRIEF - slice-03 Technical-plan review decision data
+
+## Summary
+
+Completed technical-plan candidate enrichment so approval decisions reflect review status, approval recommendation, blocking state, required fixes, optional hardening, risks, and next command.
+
+## Validation
+
+- [x] Plan-review state tests pass.
+- [x] Revise blocking test passes.
+- [x] Approve-with-risk test passes.
+
+## Pending
+
+- None.
+
+## Remaining Risks
+
+- None critical; workflow surfaces are integrated in slice-06.

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-03-technical-plan-review-decision-data/EXECUTION_BRIEF.md

@@ -0,0 +1,50 @@
+# EXECUTION_BRIEF - slice-03 Technical-plan review decision data
+
+## Context
+
+Technical-plan approval already has safety checks, but users need visible decision data before approving.
+
+## Objective
+
+Surface plan-review state and recommendation consistently during technical-plan approval.
+
+## Scope
+
+- `src/create-quiver/lib/ai/plan-review.js`
+- approval candidate helpers
+- `runApprove` selector output if needed
+- focused plan-review tests
+
+## Acceptance Criteria
+
+- Missing, stale, unapproved, approve, approve-with-risk, and revise states are distinguishable.
+- Revise blocks approval.
+- Approve-with-risk remains approvable and visible.
+- Required fixes, optional hardening, risks, and next command are shown where relevant.
+- Existing technical-plan approval validation remains intact.
+
+## Technical Plan Summary
+
+Extend the shared candidate data with normalized review-decision fields and consume them in the approval selector.
+
+## Suggested Steps
+
+1. Normalize review decision state.
+2. Add candidate annotations for technical-plan.
+3. Update selector rendering or approval summaries.
+4. Add tests for all review states.
+
+## Restrictions
+
+- Do not modify provider prompts.
+- Do not change review-result schema unless needed for backward compatibility.
+
+## Risks
+
+- Displaying unapproved vs reviewed states incorrectly can lead users to approve the wrong artifact.
+
+## Completion Checklist
+
+- [ ] Review states normalized.
+- [ ] Blocking behavior preserved.
+- [ ] Tests cover all review states.

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-03-technical-plan-review-decision-data/slice.json

@@ -0,0 +1,75 @@
+{
+  "slice_id": "slice-03-technical-plan-review-decision-data",
+  "ticket": "QUIVER-33-03",
+  "type": "feature",
+  "title": "Technical-plan review decision data",
+  "objective": "Make technical-plan approval decisions visibly reflect plan-review state and recommendation.",
+  "description": "Ensures approval selection and command output distinguish missing, stale, unapproved, approve, approve-with-risk, and revise review states without weakening existing gates.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v33-technical-plan-review-decision-data",
+    "branch_name": "feature/QUIVER-33-03-v33-technical-plan-review-decision-data"
+  },
+  "files": [
+    "src/create-quiver/lib/ai/plan-review.js",
+    "src/create-quiver/lib/approvals.js",
+    "src/create-quiver/commands/ai.js",
+    "tests/commands/ai-review-plan.test.js",
+    "tests/lib/approvals.test.js"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/lib/ai/plan-review.js",
+    "src/create-quiver/lib/approvals.js",
+    "src/create-quiver/commands/ai.js",
+    "tests/commands/ai-review-plan.test.js",
+    "SPEC.md"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/lib/ai/plan-review.js",
+    "src/create-quiver/lib/approvals.js",
+    "src/create-quiver/commands/ai.js",
+    "tests/commands/ai-review-plan.test.js",
+    "tests/lib/approvals.test.js",
+    "specs/quiver-v33-approval-ux-and-planner-progress/**"
+  ],
+  "depends_on": [
+    "slice-01-approval-candidates-model",
+    "slice-02-approve-interactive-selection"
+  ],
+  "parallel_safe": "yes",
+  "parallel_safe_reason": "Can proceed after approval selection if command-file conflicts are coordinated.",
+  "must": [
+    "Represent plan-review statuses missing, stale, unapproved, reviewed approve, reviewed approve-with-risk, and reviewed revise.",
+    "Block approval when recommendation is revise.",
+    "Allow approve-with-risk with visible risk context.",
+    "Show required fixes count and next command when blocked.",
+    "Preserve existing explicit technical-plan approval validation."
+  ],
+  "not_included": [
+    "Changing how reviewers generate review metadata.",
+    "Changing structured review schema unless required for display.",
+    "Relaxing stale review checks."
+  ],
+  "acceptance": [
+    "Technical-plan selector shows missing review as not ready.",
+    "Stale review blocks or redirects to review-plan.",
+    "Unapproved review guides to approve with the reviewed version when valid.",
+    "Approve-with-risk exposes optional hardening and risks.",
+    "Revise recommendation blocks approval.",
+    "Explicit version approval still enforces the same blocking rules."
+  ],
+  "tests": [
+    "node --test tests/commands/ai-review-plan.test.js tests/lib/approvals.test.js",
+    "git diff --check"
+  ],
+  "validation_hints": [
+    "Reuse existing `reviewBlocksApproval` semantics.",
+    "Add tests around every plan-review status."
+  ],
+  "estimated_hours": 6,
+  "status": "completed",
+  "blocked_reason": null,
+  "actual_hours": 6,
+  "completed_at": "2026-05-28"
+}

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-04-revise-input-guardrails/CLOSURE_BRIEF.md

@@ -0,0 +1,19 @@
+# CLOSURE_BRIEF - slice-04 Revise input guardrails
+
+## Summary
+
+Completed revise input guardrails for missing `--input` values, nonexistent feedback files, and accidental extra positional arguments before any provider execution.
+
+## Validation
+
+- [x] Missing input tests pass.
+- [x] Nonexistent file tests pass.
+- [x] Valid revise regression tests pass.
+
+## Pending
+
+- None.
+
+## Remaining Risks
+
+- None critical; broader CLI contract coverage passed in the full suite.

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-04-revise-input-guardrails/EXECUTION_BRIEF.md

@@ -0,0 +1,51 @@
+# EXECUTION_BRIEF - slice-04 Revise input guardrails
+
+## Context
+
+The requirement explicitly called out incomplete `ai revise --phase technical-plan --input` usage and an accidental trailing `s`. Production review expanded the scope to acceptance revise as well.
+
+## Objective
+
+Make invalid `ai revise --input` usage fail clearly before provider execution for both planner phases.
+
+## Scope
+
+- CLI argument parsing if needed.
+- `runRevise` / `runPlan` input validation.
+- focused command tests.
+
+## Acceptance Criteria
+
+- Missing `--input` value is detected for acceptance and technical-plan.
+- Nonexistent feedback files fail before provider execution.
+- Accidental extra arguments are reported or rejected clearly.
+- No-TTY/CI never prompts.
+- Valid revise commands remain unchanged.
+
+## Technical Plan Summary
+
+Add explicit guardrails at the parser/command boundary, then test command-level behavior with provider execution disabled or asserted absent.
+
+## Suggested Steps
+
+1. Inspect parse behavior for `--input` without value.
+2. Add a consistent actionable error.
+3. Add nonexistent-file regression coverage.
+4. Add accidental extra-arg coverage if parser supports it.
+5. Confirm valid revise flows still pass.
+
+## Restrictions
+
+- Do not run a provider when input is invalid.
+- Do not auto-create feedback files.
+
+## Risks
+
+- Parser changes can affect other commands that use `--input`.
+
+## Completion Checklist
+
+- [ ] Missing input covered for both phases.
+- [ ] Nonexistent input covered.
+- [ ] Extra argument behavior covered.
+- [ ] Valid revise flow still passes.

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-04-revise-input-guardrails/slice.json

@@ -0,0 +1,73 @@
+{
+  "slice_id": "slice-04-revise-input-guardrails",
+  "ticket": "QUIVER-33-04",
+  "type": "fix",
+  "title": "Revise input guardrails",
+  "objective": "Handle missing, incomplete, nonexistent, and accidental extra `ai revise --input` usage before provider execution.",
+  "description": "Hardens `ai revise` for both acceptance and technical-plan phases so incomplete inputs fail actionably or guide the user in TTY without silently ignoring mistakes.",
+  "git": {
+    "branch_type": "fix",
+    "base_branch": "main",
+    "branch_slug": "v33-revise-input-guardrails",
+    "branch_name": "fix/QUIVER-33-04-v33-revise-input-guardrails"
+  },
+  "files": [
+    "src/create-quiver/index.js",
+    "src/create-quiver/commands/ai.js",
+    "tests/commands/ai-plan.test.js",
+    "tests/commands/cli-contract.test.js"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/index.js",
+    "src/create-quiver/commands/ai.js",
+    "tests/commands/ai-plan.test.js",
+    "tests/commands/cli-contract.test.js",
+    "SPEC.md"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/index.js",
+    "src/create-quiver/commands/ai.js",
+    "tests/commands/ai-plan.test.js",
+    "tests/commands/cli-contract.test.js",
+    "specs/quiver-v33-approval-ux-and-planner-progress/**"
+  ],
+  "depends_on": [
+    "slice-01-approval-candidates-model"
+  ],
+  "parallel_safe": "yes",
+  "parallel_safe_reason": "Can run after slice-01, but coordinate command-file conflicts with slice-02 and slice-05.",
+  "must": [
+    "Detect `ai revise --phase acceptance --input` without a value.",
+    "Detect `ai revise --phase technical-plan --input` without a value.",
+    "Fail before provider execution for missing or nonexistent files.",
+    "Handle accidental trailing args such as `s` with a clear error or documented parse behavior.",
+    "In no-TTY/CI, show an explicit corrected command.",
+    "In TTY, show focused guidance or selector only if useful artifacts exist."
+  ],
+  "not_included": [
+    "Generating feedback files automatically.",
+    "Running providers when input is invalid.",
+    "Changing revise prompt semantics."
+  ],
+  "acceptance": [
+    "`ai revise --phase acceptance --input` fails before provider execution with a useful next command.",
+    "`ai revise --phase technical-plan --input` fails before provider execution with a useful next command.",
+    "Nonexistent feedback input fails before provider execution.",
+    "Accidental extra args are not silently accepted.",
+    "No-TTY/CI does not prompt.",
+    "Existing valid revise commands still work."
+  ],
+  "tests": [
+    "node --test tests/commands/ai-plan.test.js tests/commands/cli-contract.test.js",
+    "git diff --check"
+  ],
+  "validation_hints": [
+    "Cover parser-level missing value behavior and command-level missing file behavior separately.",
+    "Avoid provider mocks for invalid input paths because provider execution should not start."
+  ],
+  "estimated_hours": 6,
+  "status": "completed",
+  "blocked_reason": null,
+  "actual_hours": 6,
+  "completed_at": "2026-05-28"
+}

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-05-provider-progress-alignment/CLOSURE_BRIEF.md

@@ -0,0 +1,20 @@
+# CLOSURE_BRIEF - slice-05 Provider progress alignment
+
+## Summary
+
+Completed provider progress alignment for planner/reviewer flows. `ai plan`, `ai revise`, `ai review-plan`, and `ai repair-plan` now use real preparation checks and provider spinner behavior in human TTY mode.
+
+## Validation
+
+- [x] Planner progress tests pass.
+- [x] Review-plan progress tests pass.
+- [x] Repair-plan progress tests pass.
+- [x] Machine-output cleanliness tests pass.
+
+## Pending
+
+- None.
+
+## Remaining Risks
+
+- None critical; final smoke and package readiness passed.

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-05-provider-progress-alignment/EXECUTION_BRIEF.md

@@ -0,0 +1,51 @@
+# EXECUTION_BRIEF - slice-05 Provider progress alignment
+
+## Context
+
+`ai prepare-context --with-planner` provides the UX baseline. Planner and reviewer commands must look alive during real provider execution without affecting machine output.
+
+## Objective
+
+Align progress/loaders for `ai plan`, `ai revise`, `ai review-plan`, and `ai repair-plan`, and audit existing long-running flows.
+
+## Scope
+
+- `src/create-quiver/commands/ai.js`
+- existing CLI UX helpers if needed
+- focused progress tests
+
+## Acceptance Criteria
+
+- Live provider-backed planner/reviewer commands show progress in human TTY.
+- Dry-run and print-prompt do not show loaders.
+- JSON/no-TTY/CI/no-color stay clean.
+- Provider failure stops the spinner.
+- Existing compliant flows are protected by regression tests or noted.
+
+## Technical Plan Summary
+
+Reuse `createCommandUx`, `shouldShowHumanProgress`, `writeProgressChecks`, and `runProviderWithProgress`. Avoid new abstractions unless real duplication appears.
+
+## Suggested Steps
+
+1. Audit current progress usage.
+2. Add missing progress stages.
+3. Add tests with injected spinner/UX helpers.
+4. Add no-color/no-TTY regression coverage where gaps exist.
+
+## Restrictions
+
+- Do not change provider prompts.
+- Do not add progress to JSON output.
+- Do not mark fake work as completed.
+
+## Risks
+
+- Progress output can be hard to test if implementation is coupled to terminal rendering.
+
+## Completion Checklist
+
+- [ ] Planner progress covered.
+- [ ] Review-plan progress covered.
+- [ ] Repair-plan progress covered.
+- [ ] Machine-output cleanliness covered.

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-05-provider-progress-alignment/slice.json

@@ -0,0 +1,91 @@
+{
+  "slice_id": "slice-05-provider-progress-alignment",
+  "ticket": "QUIVER-33-05",
+  "type": "feature",
+  "title": "Provider progress alignment",
+  "objective": "Align progress messages and loaders for provider-backed planner and reviewer commands.",
+  "description": "Ensures `ai plan`, `ai revise`, `ai review-plan`, and `ai repair-plan` use the same production-safe progress contract as `ai prepare-context --with-planner`, while auditing existing long-running flows.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v33-provider-progress-alignment",
+    "branch_name": "feature/QUIVER-33-05-v33-provider-progress-alignment"
+  },
+  "files": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/executor.js",
+    "src/create-quiver/lib/ai/execution-plan.js",
+    "src/create-quiver/lib/cli/ux.js",
+    "tests/commands/ai-plan.test.js",
+    "tests/commands/ai-review-plan.test.js",
+    "tests/commands/ai-prepare-context-planner.test.js",
+    "tests/commands/ai-execute-slice.test.js",
+    "tests/commands/ai-execute-plan.test.js",
+    "tests/commands/ai-pr.test.js",
+    "tests/lib/cli-ux.test.js"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/executor.js",
+    "src/create-quiver/lib/ai/execution-plan.js",
+    "src/create-quiver/lib/cli/ux.js",
+    "tests/commands/ai-prepare-context-planner.test.js",
+    "tests/lib/cli-ux.test.js",
+    "SPEC.md"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/executor.js",
+    "src/create-quiver/lib/ai/execution-plan.js",
+    "src/create-quiver/lib/cli/ux.js",
+    "tests/commands/ai-plan.test.js",
+    "tests/commands/ai-review-plan.test.js",
+    "tests/commands/ai-prepare-context-planner.test.js",
+    "tests/commands/ai-execute-slice.test.js",
+    "tests/commands/ai-execute-plan.test.js",
+    "tests/commands/ai-pr.test.js",
+    "tests/lib/cli-ux.test.js",
+    "specs/quiver-v33-approval-ux-and-planner-progress/**"
+  ],
+  "depends_on": [
+    "slice-01-approval-candidates-model"
+  ],
+  "parallel_safe": "yes",
+  "parallel_safe_reason": "Can run after slice-01, but command-file conflicts with approval slices should be coordinated.",
+  "must": [
+    "Show human progress for `ai plan --phase acceptance` live provider execution.",
+    "Show human progress for `ai plan --phase technical-plan` live provider execution.",
+    "Show human progress for `ai revise` live provider execution through the shared plan path.",
+    "Show human progress for `ai review-plan` live provider execution.",
+    "Show human progress for `ai repair-plan` live provider execution.",
+    "Do not show loaders for dry-run or print-prompt.",
+    "Do not contaminate JSON, no-color, CI, or no-TTY output.",
+    "Audit existing long-running flows and add regression tests only where behavior should remain unchanged."
+  ],
+  "not_included": [
+    "Changing provider prompts.",
+    "Changing provider adapters.",
+    "Adding new UX dependencies."
+  ],
+  "acceptance": [
+    "Planner commands show real staged progress in human mode.",
+    "Reviewer and repair-plan commands show real staged progress in human mode.",
+    "Dry-run and print-prompt remain provider-free and loader-free.",
+    "Provider failure stops spinner and surfaces actionable context.",
+    "Existing prepare-context, onboard, execute-slice, execute-plan, and pr progress behavior is covered by regression tests or explicitly documented as already compliant."
+  ],
+  "tests": [
+    "node --test tests/commands/ai-plan.test.js tests/commands/ai-review-plan.test.js tests/commands/ai-prepare-context-planner.test.js tests/lib/cli-ux.test.js",
+    "node --test tests/commands/ai-execute-slice.test.js tests/commands/ai-execute-plan.test.js tests/commands/ai-pr.test.js",
+    "git diff --check"
+  ],
+  "validation_hints": [
+    "Use injected spinner helpers in tests.",
+    "Avoid fake progress checks that claim completed work before it happens."
+  ],
+  "estimated_hours": 8,
+  "status": "completed",
+  "blocked_reason": null,
+  "actual_hours": 8,
+  "completed_at": "2026-05-28"
+}

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-06-workflow-surface-integration/CLOSURE_BRIEF.md

@@ -0,0 +1,20 @@
+# CLOSURE_BRIEF - slice-06 Workflow surface integration
+
+## Summary
+
+Completed workflow surface integration through shared approval-candidate helpers. `ai approvals`, `flow`, `ai status`, `ai resume`, and `spec create --interactive` now align on recommended approval versions and blocked review guidance.
+
+## Validation
+
+- [x] Approval status tests pass.
+- [x] Flow tests pass.
+- [x] Run-state tests pass.
+- [x] Spec create tests pass.
+
+## Pending
+
+- None.
+
+## Remaining Risks
+
+- None critical; drift-prone surfaces are covered by focused tests and full suite.

package/specs/quiver-v33-approval-ux-and-planner-progress/slices/slice-06-workflow-surface-integration/EXECUTION_BRIEF.md

@@ -0,0 +1,53 @@
+# EXECUTION_BRIEF - slice-06 Workflow surface integration
+
+## Context
+
+After approval selection and review-decision data exist, other workflow surfaces must consume the same source of truth to prevent contradictory next steps.
+
+## Objective
+
+Integrate shared approval candidates into `ai approvals`, `flow`, `ai status`, `ai resume`, and `spec create --interactive`.
+
+## Scope
+
+- `src/create-quiver/commands/ai.js`
+- `src/create-quiver/commands/flow.js`
+- `src/create-quiver/commands/spec.js`
+- `src/create-quiver/lib/ai/run-state.js`
+- focused tests
+
+## Acceptance Criteria
+
+- `ai approvals` and `ai approve` recommend the same current version.
+- `flow` points to safe next steps consistent with candidates.
+- `ai status` and `ai resume` do not contradict approval guidance.
+- `spec create --interactive` shows reviewed and approved technical-plan input clearly.
+- Read-only outputs remain prompt-free.
+
+## Technical Plan Summary
+
+Replace duplicated next-command decisions with the shared candidate model where approval phases are involved.
+
+## Suggested Steps
+
+1. Update approval status output.
+2. Update flow decision points.
+3. Update run-state next-command guidance.
+4. Update spec create interactive summary.
+5. Add drift-prevention tests.
+
+## Restrictions
+
+- Do not prompt from read-only commands.
+- Do not change spec generation semantics.
+
+## Risks
+
+- Existing snapshot-style output tests may need precise updates.
+
+## Completion Checklist
+
+- [ ] `ai approvals` integrated.
+- [ ] `flow` integrated.
+- [ ] `ai status/resume` integrated.
+- [ ] `spec create --interactive` integrated.