create-quiver 0.17.1 → 0.17.3

package/specs/quiver-v55-analyze-project-doc-apply-ux/slices/slice-03-noninteractive-apply-engine/pr.md

@@ -0,0 +1,132 @@
+## Title
+
+QUIVER-55-03 - Non-Interactive Analyze Project Apply Engine
+
+## Summary
+
+Implements the third v55 slice for `ai analyze-project`: `--apply-docs --yes` now applies validated documentation proposals through a safe non-interactive write engine.
+
+## PR Policy
+
+- One slice: `slice-03-noninteractive-apply-engine`.
+- Runtime behavior change is limited to the non-interactive `--apply-docs --yes` path.
+- This PR does not implement interactive selector UX or saved proposal application by run id.
+- Human merge is required because this PR changes runtime write behavior.
+
+## Scope
+
+Included:
+
+- Persist proposal artifacts before applying docs.
+- Build the write plan from a normalized validated proposal.
+- Block dirty target docs in `--yes` unless `--allow-dirty-docs` is passed.
+- Block stale target docs when current hashes no longer match the saved proposal manifest.
+- Create snapshots before final doc writes.
+- Write approved docs only through the existing managed-block path.
+- Save `.quiver/runs/<run-id>/writes/analyze-project-doc-writes.json`.
+- Run post-write validation.
+- Support `--apply-docs --yes --json`.
+- Preserve existing `--review` behavior.
+
+Excluded:
+
+- No interactive selector for plain `--apply-docs`.
+- No `ai analyze-project apply --run <run-id>`.
+- No live provider smoke in CI.
+
+## Files
+
+- `src/create-quiver/commands/ai.js`
+- `src/create-quiver/lib/ai/analyze-project-apply.js`
+- `src/create-quiver/lib/ai/analyze-project-proposal.js`
+- `tests/commands/ai-analyze-project-review.test.js`
+- `tests/lib/ai-analyze-project-apply.test.js`
+- `tests/lib/ai-analyze-project-proposal.test.js`
+- `specs/quiver-v55-analyze-project-doc-apply-ux/**`
+
+## How to Test (DETAILED - REQUIRED)
+
+### Required Environment
+
+- Node.js supported by this repository.
+- npm dependencies installed.
+- No provider credentials are required for automated tests.
+
+### Worktree Access
+
+```bash
+git checkout feature/QUIVER-55-03-v55-noninteractive-apply-engine
+```
+
+### Run the Project
+
+No dev server is required. This is a CLI docs-apply engine slice.
+
+### Use Cases
+
+#### Case 1: Apply valid docs non-interactively
+
+```bash
+node --test tests/commands/ai-analyze-project-review.test.js
+```
+
+Expected: `--apply-docs --yes` writes approved docs, saves proposal artifacts, creates a snapshot, saves the write manifest, and passes post-write validation.
+
+#### Case 2: Automation JSON output
+
+```bash
+node --test tests/commands/ai-analyze-project-review.test.js
+```
+
+Expected: `--apply-docs --yes --json` emits parseable JSON containing run id, proposal artifacts, write plan, snapshot, written docs, post-write validation, and write manifest.
+
+#### Case 3: Dirty docs are blocked by default
+
+```bash
+node --test tests/commands/ai-analyze-project-review.test.js
+```
+
+Expected: existing target docs block `--yes` unless `--allow-dirty-docs` is provided; final docs remain unchanged when blocked.
+
+#### Case 4: Stale docs are blocked
+
+```bash
+node --test tests/lib/ai-analyze-project-apply.test.js
+```
+
+Expected: hash mismatches between proposal manifest and current target docs fail before writing final docs.
+
+### Technical Verification
+
+```bash
+node --test tests/commands/ai-analyze-project-review.test.js tests/lib/ai-analyze-project-apply.test.js tests/lib/ai-analyze-project-proposal.test.js tests/lib/ai-analyze-project-validation.test.js
+node --test tests/commands/ai-analyze-project-provider.test.js tests/commands/ai-analyze-project.test.js tests/commands/cli-contract.test.js tests/commands/ux-flags.test.js
+node --test tests/lib/ai/analyze-project-repair.test.js
+npm run docs:check
+npm run schema:slice:check
+node bin/create-quiver.js spec validate specs/quiver-v55-analyze-project-doc-apply-ux --strict
+git diff --check
+npm test
+```
+
+## Evidence
+
+- All technical verification commands passed locally.
+- Full `npm test` passed locally: 730 tests.
+- `slice-03` closure brief records executed evidence.
+
+## Rollback
+
+Revert this PR:
+
+```bash
+git revert <merge-commit-sha>
+```
+
+This disables non-interactive docs apply while preserving the earlier CLI contract and `--save-proposal` artifacts.
+
+## Risks / Notes
+
+- `--apply-docs` without `--yes` still intentionally blocks until the interactive selector slice lands.
+- Existing docs are treated as dirty targets and require `--allow-dirty-docs` for automation.
+- Saved proposal application remains a later slice.

package/specs/quiver-v55-analyze-project-doc-apply-ux/slices/slice-03-noninteractive-apply-engine/slice.json

@@ -0,0 +1,76 @@
+{
+  "slice_id": "slice-03-noninteractive-apply-engine",
+  "ticket": "QUIVER-55-03",
+  "type": "feature",
+  "title": "Non-Interactive Apply Engine",
+  "objective": "Apply validated analyze-project doc proposals safely through --apply-docs --yes without selector UI.",
+  "description": "Adds the reusable apply engine: proposal persistence, dirty/stale preflight, snapshot, doc writes, write manifest, post-write validation, and JSON output.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v55-noninteractive-apply-engine",
+    "branch_name": "feature/QUIVER-55-03-v55-noninteractive-apply-engine"
+  },
+  "files": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/analyze-project-apply.js",
+    "src/create-quiver/lib/ai/analyze-project-proposal.js",
+    "tests/commands/ai-analyze-project-review.test.js",
+    "tests/lib/ai-analyze-project-apply.test.js",
+    "tests/lib/ai-analyze-project-proposal.test.js",
+    "tests/lib/ai-analyze-project-validation.test.js",
+    "specs/quiver-v55-analyze-project-doc-apply-ux/**"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/lib/ai/analyze-project-docs.js",
+    "src/create-quiver/lib/ai/analyze-project-validation.js",
+    "src/create-quiver/lib/ai/analyze-project-review.js"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/analyze-project-apply.js",
+    "src/create-quiver/lib/ai/analyze-project-proposal.js",
+    "tests/commands/ai-analyze-project-review.test.js",
+    "tests/lib/ai-analyze-project-apply.test.js",
+    "tests/lib/ai-analyze-project-proposal.test.js",
+    "tests/lib/ai-analyze-project-validation.test.js",
+    "specs/quiver-v55-analyze-project-doc-apply-ux/**"
+  ],
+  "depends_on": [
+    "slice-01-cli-proposal-contract",
+    "slice-02-save-proposal-flow"
+  ],
+  "parallel_safe": "never",
+  "parallel_safe_reason": "This slice owns the docs write boundary and must be reviewed independently.",
+  "must": [
+    "Persist proposal before apply.",
+    "Create snapshot before final doc writes.",
+    "Save write manifest.",
+    "Run post-write validation."
+  ],
+  "not_included": [
+    "Interactive selector.",
+    "Apply saved run.",
+    "Docs release smoke."
+  ],
+  "acceptance": [
+    "--apply-docs --yes writes valid allowed docs.",
+    "Dirty docs block --yes unless --allow-dirty-docs is passed.",
+    "Invalid proposal writes no final docs.",
+    "--apply-docs --yes --json is parseable and complete."
+  ],
+  "tests": [
+    "node --test tests/commands/ai-analyze-project-review.test.js",
+    "node --test tests/lib/ai-analyze-project-apply.test.js",
+    "node --test tests/lib/ai-analyze-project-proposal.test.js",
+    "node --test tests/lib/ai-analyze-project-validation.test.js",
+    "node bin/create-quiver.js spec validate specs/quiver-v55-analyze-project-doc-apply-ux --strict",
+    "git diff --check"
+  ],
+  "estimated_hours": 6,
+  "actual_hours": 6,
+  "status": "completed",
+  "blocked_reason": null,
+  "started_at": "2026-06-12T00:00:00.000Z",
+  "completed_at": "2026-06-12T00:00:00.000Z"
+}

package/specs/quiver-v55-analyze-project-doc-apply-ux/slices/slice-04-interactive-apply-ux/CLOSURE_BRIEF.md

@@ -0,0 +1,47 @@
+# CLOSURE_BRIEF - slice-04 interactive apply UX
+
+## Summary
+
+Completed. Plain TTY `ai analyze-project --deep --apply-docs` now runs the provider, builds the validated proposal, and shows an explained selector instead of forcing a large editor buffer. The selector routes apply/save/edit/cancel/view-diff actions through the existing safe proposal, review, and apply flows.
+
+## Behavior Delivered
+
+- TTY `--apply-docs` shows compact summary, proposed files, changed/dirty counts, selected context artifact, and option descriptions.
+- Dynamic recommendation puts `Apply documentation` first for clean creates and `View diff` first for dirty/update targets.
+- `View diff` saves the full diff artifact, prints only a bounded terminal preview, and requires a second decision.
+- `Save proposal` writes `.quiver/runs/<run-id>/proposal/*` artifacts only.
+- `Edit proposal` reuses the existing review/editor/confirmation flow.
+- `Cancel` writes no final docs and records `apply-canceled` status.
+- no-TTY or CI `--apply-docs` without `--yes` fails before provider execution with actionable guidance.
+- English and Spanish selector copy are covered.
+
+## Required Evidence
+
+- Apply option test passes.
+- Diff option test passes.
+- Save option test passes.
+- Edit option test passes.
+- Cancel option test passes.
+- no-TTY error test passes.
+
+## Validation
+
+Executed:
+
+```bash
+node --test tests/commands/ai-analyze-project-review.test.js
+node --test tests/lib/cli-ux.test.js
+node --test tests/commands/ai-analyze-project.test.js
+node --test tests/commands/cli-contract.test.js
+node --test tests/lib/i18n-catalog.test.js
+node --test tests/commands/ux-flags.test.js
+node --test tests/commands/ai-analyze-project-provider.test.js
+node --test tests/lib/ai-analyze-project-validation.test.js tests/lib/ai/analyze-project-repair.test.js
+npm run docs:check
+npm run schema:slice:check
+node bin/create-quiver.js spec validate specs/quiver-v55-analyze-project-doc-apply-ux --strict
+git diff --check
+npm test
+```
+
+Result: all passed. Full `npm test` passed with 741 tests.

package/specs/quiver-v55-analyze-project-doc-apply-ux/slices/slice-04-interactive-apply-ux/EXECUTION_BRIEF.md

@@ -0,0 +1,70 @@
+# EXECUTION_BRIEF - slice-04 interactive apply UX
+
+## Context
+
+The main user-facing improvement is an explained selector that avoids forcing users into a large editor buffer. This selector must orchestrate already-tested save/apply/edit actions instead of owning write logic.
+
+## Objective
+
+Implement `--apply-docs` TTY selector UX with English/Spanish messages and safe action routing.
+
+## Scope
+
+- Show compact analysis/proposal summary.
+- Show proposed files and action counts.
+- Show repair summary and artifact paths.
+- Render selector options with one-line descriptions.
+- Dynamic recommendation:
+  - Apply documentation when target docs are clean/low-risk;
+  - View diff when docs are existing/dirty/high-risk.
+- Implement actions:
+  - apply;
+  - view diff then apply/save/cancel;
+  - save proposal;
+  - edit proposal via existing `--review` path;
+  - cancel.
+- Truncate human diff and save full diff artifact.
+
+## Acceptance Criteria
+
+- TTY `--apply-docs` shows explained options.
+- no-TTY `--apply-docs` without `--yes` fails with actionable message.
+- Cancel writes no final docs.
+- Save proposal from selector writes artifacts only.
+- Edit proposal reuses editor flow and handles editor failure by returning actionable guidance.
+- View diff does not accidentally apply docs and requires a second decision.
+- `--review` remains backward compatible.
+
+## Expected Files To Modify
+
+- `src/create-quiver/commands/ai.js`
+- `src/create-quiver/lib/ai/analyze-project-interactive.js`
+- `src/create-quiver/lib/ai/analyze-project-apply.js`
+- `src/create-quiver/lib/i18n/messages/en.js`
+- `src/create-quiver/lib/i18n/messages/es.js`
+- `tests/commands/ai-analyze-project-review.test.js`
+- `tests/lib/cli-ux.test.js`
+- this slice closure/status/evidence files
+
+## Validation Required
+
+```bash
+node --test tests/commands/ai-analyze-project-review.test.js
+node --test tests/lib/cli-ux.test.js
+node bin/create-quiver.js spec validate specs/quiver-v55-analyze-project-doc-apply-ux --strict
+git diff --check
+```
+
+## Completion Checklist
+
+- TTY selector presents explained choices in the configured language.
+- no-TTY apply without `--yes` fails with actionable guidance.
+- Diff, save, edit, cancel, and apply paths are tested.
+- Terminal diff output remains bounded.
+- Slice closure brief updated with evidence.
+
+## Constraints
+
+- Do not duplicate write logic inside selector code.
+- Do not make `--interactive` required for `--apply-docs`.
+- Do not print unbounded diffs in terminal.

package/specs/quiver-v55-analyze-project-doc-apply-ux/slices/slice-04-interactive-apply-ux/pr.md

@@ -0,0 +1,141 @@
+## Title
+
+QUIVER-55-04 - Interactive Analyze Project Apply UX
+
+## Summary
+
+Implements the fourth v55 slice for `ai analyze-project`: plain TTY `--apply-docs` now shows an explained selector after provider validation instead of forcing users into a large editor buffer.
+
+The selector supports:
+
+- Apply documentation.
+- View bounded diff, then choose again.
+- Save proposal artifacts only.
+- Edit proposal through the existing review flow.
+- Cancel without final docs writes.
+
+## PR Policy
+
+- One slice, one commit, one PR.
+- Runtime UX change, so this must be an individual PR.
+- Human merge required after CI passes.
+
+## Scope
+
+Included:
+
+- Interactive selector module for `--apply-docs`.
+- Safe routing from selector actions into existing save/apply/review flows.
+- Dynamic recommendation based on clean creates vs dirty/update targets.
+- Bounded terminal diff with full diff artifact saved under `.quiver/runs`.
+- English and Spanish selector copy with option descriptions.
+- Tests for apply, save, edit, cancel, view-diff second decision, no-TTY guard, and selector recommendation.
+- Slice status, evidence, and closure updates.
+
+Excluded:
+
+- `ai analyze-project apply --run <run-id>`.
+- Release smoke docs.
+- Product repo code changes.
+
+## Files
+
+- `src/create-quiver/commands/ai.js`
+- `src/create-quiver/lib/ai/analyze-project-interactive.js`
+- `src/create-quiver/lib/i18n/messages/en.js`
+- `src/create-quiver/lib/i18n/messages/es.js`
+- `tests/commands/ai-analyze-project-review.test.js`
+- `tests/lib/cli-ux.test.js`
+- `specs/quiver-v55-analyze-project-doc-apply-ux/**`
+
+## How to Test (DETAILED - REQUIRED)
+
+### Required Environment
+
+- Node/npm environment for this repo.
+- No live provider credentials required for automated tests.
+
+### Worktree Access
+
+```bash
+git checkout feature/QUIVER-55-04-v55-interactive-apply-ux
+```
+
+### Run the Project
+
+No dev server is required. This is CLI behavior.
+
+### Use Cases
+
+1. Interactive apply selector:
+
+```bash
+node --test tests/commands/ai-analyze-project-review.test.js
+```
+
+Expected: selector paths for apply, save, edit, cancel, view diff, no-TTY, and Spanish copy pass.
+
+2. UX helper behavior:
+
+```bash
+node --test tests/lib/cli-ux.test.js
+```
+
+Expected: selector recommendations and bounded diff formatting pass.
+
+3. Contract regressions:
+
+```bash
+node --test tests/commands/ai-analyze-project.test.js
+node --test tests/commands/cli-contract.test.js
+node --test tests/commands/ux-flags.test.js
+node --test tests/lib/i18n-catalog.test.js
+```
+
+Expected: CLI contracts, help, UX flags, and i18n catalog remain valid.
+
+### Technical Verification
+
+```bash
+node --test tests/commands/ai-analyze-project-provider.test.js
+node --test tests/lib/ai-analyze-project-validation.test.js tests/lib/ai/analyze-project-repair.test.js
+npm run docs:check
+npm run schema:slice:check
+node bin/create-quiver.js spec validate specs/quiver-v55-analyze-project-doc-apply-ux --strict
+git diff --check
+npm test
+```
+
+Expected: all pass.
+
+## Evidence
+
+Executed locally:
+
+```bash
+node --test tests/commands/ai-analyze-project-review.test.js
+node --test tests/lib/cli-ux.test.js
+node --test tests/commands/ai-analyze-project.test.js
+node --test tests/commands/cli-contract.test.js
+node --test tests/lib/i18n-catalog.test.js
+node --test tests/commands/ux-flags.test.js
+node --test tests/commands/ai-analyze-project-provider.test.js
+node --test tests/lib/ai-analyze-project-validation.test.js tests/lib/ai/analyze-project-repair.test.js
+npm run docs:check
+npm run schema:slice:check
+node bin/create-quiver.js spec validate specs/quiver-v55-analyze-project-doc-apply-ux --strict
+git diff --check
+npm test
+```
+
+Result: all passed. Full `npm test` passed with 741 tests.
+
+## Rollback
+
+Revert this PR. `--apply-docs --yes`, `--save-proposal`, and `--review` remain available from earlier slices if this selector UX must be removed.
+
+## Risks / Notes
+
+- `--apply-docs` without `--yes` now runs the provider in interactive TTY mode before showing the selector.
+- no-TTY and CI usage still require `--yes`; otherwise the command fails before provider execution.
+- `apply --run <run-id>` remains intentionally deferred to slice-05.

package/specs/quiver-v55-analyze-project-doc-apply-ux/slices/slice-04-interactive-apply-ux/slice.json

@@ -0,0 +1,72 @@
+{
+  "slice_id": "slice-04-interactive-apply-ux",
+  "ticket": "QUIVER-55-04",
+  "type": "ux",
+  "title": "Interactive Apply UX",
+  "objective": "Add the explained TTY selector for --apply-docs while delegating all writes to the tested apply engine.",
+  "description": "Implements the human selector with option descriptions, dynamic recommendation, diff flow, save/edit/cancel actions, and bounded terminal output.",
+  "git": {
+    "branch_type": "feature",
+    "base_branch": "main",
+    "branch_slug": "v55-interactive-apply-ux",
+    "branch_name": "feature/QUIVER-55-04-v55-interactive-apply-ux"
+  },
+  "files": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/analyze-project-interactive.js",
+    "src/create-quiver/lib/ai/analyze-project-apply.js",
+    "src/create-quiver/lib/i18n/messages/en.js",
+    "src/create-quiver/lib/i18n/messages/es.js",
+    "tests/commands/ai-analyze-project-review.test.js",
+    "tests/lib/cli-ux.test.js",
+    "specs/quiver-v55-analyze-project-doc-apply-ux/**"
+  ],
+  "expected_read_paths": [
+    "src/create-quiver/lib/cli/ux.js",
+    "src/create-quiver/lib/ai/analyze-project-review.js",
+    "src/create-quiver/lib/ai/analyze-project-apply.js"
+  ],
+  "allowed_write_paths": [
+    "src/create-quiver/commands/ai.js",
+    "src/create-quiver/lib/ai/analyze-project-interactive.js",
+    "src/create-quiver/lib/ai/analyze-project-apply.js",
+    "src/create-quiver/lib/i18n/messages/en.js",
+    "src/create-quiver/lib/i18n/messages/es.js",
+    "tests/commands/ai-analyze-project-review.test.js",
+    "tests/lib/cli-ux.test.js",
+    "specs/quiver-v55-analyze-project-doc-apply-ux/**"
+  ],
+  "depends_on": [
+    "slice-03-noninteractive-apply-engine"
+  ],
+  "parallel_safe": "never",
+  "parallel_safe_reason": "Selector behavior depends on the completed apply engine contract.",
+  "must": [
+    "Show option descriptions.",
+    "Use dynamic recommendation.",
+    "Keep terminal diff bounded.",
+    "Route write actions through apply engine."
+  ],
+  "not_included": [
+    "Apply saved run.",
+    "Release smoke docs."
+  ],
+  "acceptance": [
+    "TTY --apply-docs shows explained selector.",
+    "no-TTY --apply-docs without --yes fails actionably.",
+    "Cancel writes no final docs.",
+    "View diff requires a second decision."
+  ],
+  "tests": [
+    "node --test tests/commands/ai-analyze-project-review.test.js",
+    "node --test tests/lib/cli-ux.test.js",
+    "node bin/create-quiver.js spec validate specs/quiver-v55-analyze-project-doc-apply-ux --strict",
+    "git diff --check"
+  ],
+  "estimated_hours": 6,
+  "actual_hours": 6,
+  "status": "completed",
+  "blocked_reason": null,
+  "started_at": "2026-06-12T00:00:00.000Z",
+  "completed_at": "2026-06-12T00:00:00.000Z"
+}

package/specs/quiver-v55-analyze-project-doc-apply-ux/slices/slice-05-apply-saved-proposal/CLOSURE_BRIEF.md

@@ -0,0 +1,42 @@
+# CLOSURE_BRIEF - slice-05 apply saved proposal
+
+## Summary
+
+Completed. `ai analyze-project apply --run <run-id>` now loads a saved analyze-project doc proposal from `.quiver/runs/<run-id>/proposal/`, revalidates the manifest and proposal JSON, and applies it through the same safe apply engine without executing a provider or requiring provider/model flags.
+
+## Behavior Delivered
+
+- Explicit run ids are validated as safe single path segments before artifact access.
+- Missing or invalid saved proposal manifests fail before docs writes.
+- Saved proposal JSON is reparsed through the doc proposal schema and allowed-path checks.
+- Saved proposal manifest doc paths must match the proposal target docs.
+- Manual proposal content edits are accepted only after revalidation and recorded as `saved-proposal-edited` in the write manifest.
+- Dirty target docs block by default and can be applied only with `--allow-dirty-docs`.
+- Stale target docs block before any final docs write.
+- `--run latest` is resolved only in an interactive terminal, and is rejected with `--yes` or `--json`.
+- Apply-run reports set `provider_execution: skipped` and do not call provider/model resolution.
+
+## Required Evidence
+
+- Apply saved run success test passes.
+- Invalid run id test passes.
+- Stale/dirty blocking tests pass.
+- No-provider-execution test passes.
+- Manual proposal edit audit test passes.
+
+## Validation
+
+Executed:
+
+```bash
+node --test tests/commands/ai-analyze-project-review.test.js
+node --test tests/lib/ai-analyze-project-proposal.test.js
+node --test tests/commands/ai-analyze-project.test.js
+npm run docs:check
+npm run schema:slice:check
+node bin/create-quiver.js spec validate specs/quiver-v55-analyze-project-doc-apply-ux --strict
+git diff --check
+npm test
+```
+
+Result: all passed. Full `npm test` passed with 747 tests.

package/specs/quiver-v55-analyze-project-doc-apply-ux/slices/slice-05-apply-saved-proposal/EXECUTION_BRIEF.md

@@ -0,0 +1,61 @@
+# EXECUTION_BRIEF - slice-05 apply saved proposal
+
+## Context
+
+Saving a proposal is only useful if users can later apply it without re-running the provider. This flow has a higher trust boundary because proposal artifacts may be stale or manually edited.
+
+## Objective
+
+Implement `ai analyze-project apply --run <run-id>` to apply saved proposals without provider execution.
+
+## Scope
+
+- Resolve and validate run id.
+- Load proposal manifest and normalized JSON.
+- Revalidate proposal schema and allowed paths.
+- Recalculate target doc hashes.
+- Detect stale/dirty docs before writing.
+- Apply through the same apply engine from slice-03.
+- Support `--yes`, `--strict`, `--allow-dirty-docs`, and `--json` where safe.
+- Optionally support `--run latest` only in TTY and never with `--yes`.
+
+## Acceptance Criteria
+
+- `apply --run <run-id>` does not execute provider and does not require provider/model.
+- Invalid or missing run id fails with actionable guidance.
+- Missing/invalid proposal manifest fails without docs writes.
+- Stale or dirty docs block non-interactive apply unless explicitly allowed.
+- Manual proposal edits are accepted only after full revalidation and recorded as edited.
+- `--run latest --yes` is rejected.
+
+## Expected Files To Modify
+
+- `src/create-quiver/commands/ai.js`
+- `src/create-quiver/lib/ai/analyze-project-apply.js`
+- `src/create-quiver/lib/ai/analyze-project-proposal.js`
+- `tests/commands/ai-analyze-project-review.test.js`
+- `tests/lib/ai-analyze-project-proposal.test.js`
+- this slice closure/status/evidence files
+
+## Validation Required
+
+```bash
+node --test tests/commands/ai-analyze-project-review.test.js
+node --test tests/lib/ai-analyze-project-proposal.test.js
+node bin/create-quiver.js spec validate specs/quiver-v55-analyze-project-doc-apply-ux --strict
+git diff --check
+```
+
+## Completion Checklist
+
+- Saved proposal apply path never executes provider.
+- Proposal manifest and normalized JSON are revalidated before writes.
+- Stale/dirty target docs behavior is tested.
+- Edited proposal audit metadata is recorded.
+- Slice closure brief updated with evidence.
+
+## Constraints
+
+- Do not execute provider in `apply --run`.
+- Do not trust saved artifacts without revalidation.
+- Do not allow ambiguous `latest` in automation.