prizmkit 1.1.189 → 1.1.190

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -1,5 +1,5 @@
1
1
  {
2
- "frameworkVersion": "1.1.189",
3
- "bundledAt": "2026-11-03T16:36:50.626Z",
4
- "bundledFrom": "0d956b2"
2
+ "frameworkVersion": "1.1.190",
3
+ "bundledAt": "2026-11-15T16:12:31.340Z",
4
+ "bundledFrom": "2c99a65"
5
5
  }
@@ -1,5 +1,5 @@
1
1
  {
2
- "version": "1.1.189",
2
+ "version": "1.1.190",
3
3
  "skills": {
4
4
  "prizmkit": {
5
5
  "description": "Framework introduction and navigation for the formal single-requirement lifecycle, project initialization, Prizm docs, and independent deployment.",
@@ -51,7 +51,7 @@ After planning is complete, you MUST:
51
51
 
52
52
  ## Draft Lifecycle (Hard Rule)
53
53
 
54
- In an ordinary successful flow, `.prizmkit/plans/bug-fix-list.draft.json` is internal writer/review input, not a user approval artifact or separate deliverable. Keep it through `generate`/`append`, validation, and the complete Local Generated-Plan Review fix loop. After the Local Generated-Plan Review returns `PASS` and final validation passes, delete the ordinary internal draft before final handoff. Report `Draft cleanup: removed`. Present the final written result only after cleanup succeeds or a retained recovery reason is surfaced. Retain the draft only for explicit draft-save, interruption or blocking before terminal gates, or cleanup failure; report its path and reason. Cleanup failure does not invalidate an already reviewed and validated final list, but makes the file stale recovery state. Never delete explicit recovery state merely because an older valid final exists. For later adjustments, recreate only the minimum source draft from confirmed in-memory entries or exact-ID `show`; ordinary appends must preserve the exact current suffix and use `append --replace-tail` without loading unrelated history.
54
+ In an ordinary successful flow, `.prizmkit/plans/bug-fix-list.draft.json` is a private reasoning workspace: never presented as a deliverable, never given its own approval round, never retained after the batch lands. Build the batch there, validate it there, and run the entire Local Generated-Plan Review against it; the final list is written exactly once, after that review returns `PASS`, by a single `append` (existing list) or `generate` (new list). Then delete the draft immediately and report `Draft cleanup: removed`. A batch lost before it lands is regenerated from the confirmed decisions rather than recovered. Conditions that must keep a draft — explicit draft-save, interruption, a writer-rejected insert, cleanup failure — are defined in `${SKILL_DIR}/references/error-recovery.md`, and any retained path is named in the terminal report. For later adjustments, recreate only the minimum source draft from confirmed in-memory entries or exact-ID `show`; modifying entries this session did not create is the one case that still uses `append --replace-tail` for the exact current suffix.
55
55
 
56
56
  ## Fresh-Session Task Contract (Hard Rule)
57
57
 
@@ -165,8 +165,8 @@ For any content-generating flow that appends to an existing `.prizmkit/plans/bug
165
165
  2. Inspect those pairs only for material continuation, duplication, conflict, a shared public behavior contract, or an execution prerequisite. Add no historical content by default. Chronology, broad module overlap, shared technology, reusable implementation ideas, or similar wording alone do not establish relevance.
166
166
  3. Incorporate only the minimum exact information whose omission would make scope, behavior, compatibility, acceptance criteria, or dependency ordering incorrect. Reject vague `related to B-123` text and whole-description copying; add an old ID to `dependencies` only when it is a real execution prerequisite.
167
167
  4. Only for an explicitly identified old ID or a validation error, run `python3 ${SKILL_DIR}/scripts/validate-bug-list.py show --input .prizmkit/plans/bug-fix-list.json --id B-NNN`.
168
- 5. For append-only planning, write `.prizmkit/plans/bug-fix-list.draft.json` with exactly one root field, `bugs`, containing non-empty unchanged-schema items. This bounded draft is internal transport/recovery state, not a user approval artifact. Required draft fields are `id`, `title`, `description`, `severity`, `error_source`, `verification_type`, and `acceptance_criteria`; `status` may be omitted and defaults to `pending`; existing optional fields remain allowed. Copy no root metadata or historical bugs.
169
- 6. After consequential decisions and the pre-generation gate pass, immediately run `append`: `python3 ${SKILL_DIR}/scripts/validate-bug-list.py append --input .prizmkit/plans/bug-fix-list.draft.json --output .prizmkit/plans/bug-fix-list.json`. Do not ask for separate user approval merely to promote the draft into the final list. The command validates the base, IDs, dependencies, and complete DAG before atomic replacement. Apply review fixes only to the bounded additions and rerun `append --replace-tail` for the exact current suffix, then follow §Draft Lifecycle before presenting the final written result. `generate` remains the new-list/full-rewrite path.
168
+ 5. For append-only planning, write `.prizmkit/plans/bug-fix-list.draft.json` with exactly one root field, `bugs`, containing non-empty unchanged-schema items. This bounded draft is a private reasoning workspace and the review target, not a user approval artifact or a deliverable. Required draft fields are `id`, `title`, `description`, `severity`, `error_source`, `verification_type`, and `acceptance_criteria`; `status` may be omitted and defaults to `pending`; existing optional fields remain allowed. Copy no root metadata or historical bugs.
169
+ 6. After consequential decisions and the pre-generation gate pass, write the batch to the additions-only draft with the `context` IDs and check merge feasibility without writing: `python3 ${SKILL_DIR}/scripts/validate-bug-list.py append --input .prizmkit/plans/bug-fix-list.draft.json --output .prizmkit/plans/bug-fix-list.json --dry-run`. Do not ask for separate user approval merely to promote the draft into the final list. The dry run applies the same base, exact-ID, dependency, and complete-DAG checks the real insert will, so a `PASS` review is a promise the insert keeps. Run the real `append` once, after §Local Generated-Plan Review Gate returns `PASS`, then follow §Draft Lifecycle. `generate` remains the new-list/full-rewrite path.
170
170
  ## Scenario Routing
171
171
 
172
172
  Classify user intent first:
@@ -333,34 +333,31 @@ For every `NEEDS_FIXES` result, ask only for the consequential missing decision,
333
333
 
334
334
  ### Phase 5: Generate & Validate
335
335
 
336
- 1. **Write draft JSON**: For a new list, write the complete `.prizmkit/plans/bug-fix-list.draft.json` conforming to the final schema. For an existing list, follow §Bounded Existing-List Context and write only the additions-only `bugs` object as internal transport/recovery state.
337
- 2. **Generate or append immediately and validate** — once required semantic gates pass, invoke the matching writer without a second draft-approval pause:
338
- - New list or explicit full-list rewrite:
339
- ```bash
340
- python3 ${SKILL_DIR}/scripts/validate-bug-list.py generate --input .prizmkit/plans/bug-fix-list.draft.json --output .prizmkit/plans/bug-fix-list.json
341
- ```
336
+ 1. **Write draft JSON**: For a new list, write the complete `.prizmkit/plans/bug-fix-list.draft.json` conforming to the final schema. For an existing list, follow §Bounded Existing-List Context and write only the additions-only `bugs` object. The draft is the review target, so nothing is written to the final list yet.
337
+ 2. **Check merge feasibility without writing** — once required semantic gates pass, run the matching check without a second draft-approval pause:
338
+ - New list or explicit full-list rewrite: the draft is complete and nothing exists to overwrite, so its post-review `generate` is the validator.
342
339
  - Ordinary existing-list append:
343
340
  ```bash
344
- python3 ${SKILL_DIR}/scripts/validate-bug-list.py append --input .prizmkit/plans/bug-fix-list.draft.json --output .prizmkit/plans/bug-fix-list.json
341
+ python3 ${SKILL_DIR}/scripts/validate-bug-list.py append --input .prizmkit/plans/bug-fix-list.draft.json --output .prizmkit/plans/bug-fix-list.json --dry-run
345
342
  ```
346
- `generate` fills root defaults for a complete draft. `append` preserves root metadata and historical bugs, defaults missing new statuses, and writes only after the merged full list validates.
343
+ The dry run applies the same base, exact-ID, dependency, and complete-DAG checks the real insert will, including `append`'s preservation of root metadata and historical bugs.
347
344
  3. **If validation fails**: Fix the draft and retry (max 3 attempts). If the script is unavailable, use the checklist in `${SKILL_DIR}/references/schema-validation.md`.
348
- 4. **Run the mandatory Local Generated-Plan Review Gate** (see §Local Generated-Plan Review Gate below): load `${SKILL_DIR}/references/generated-plan-review.md` and review only newly added/changed bug entries after validation passes and before success output.
349
- 5. **Complete cleanup and output**: After local review returns `PASS` and final validation passes, follow §Draft Lifecycle, then report the file path, summary, local review, validation, cleanup result, and next steps.
345
+ 4. **Run the mandatory Local Generated-Plan Review Gate** (see §Local Generated-Plan Review Gate below): load `${SKILL_DIR}/references/generated-plan-review.md` and review only the draft batch before the insert.
346
+ 5. **Insert once and output**: After local review returns `PASS`, run a single `append`/`generate`, confirm its `appended_ids`, follow §Draft Lifecycle, then report the file path, summary, local review, inserted IDs, validation, cleanup result, and next steps.
350
347
 
351
- **Gate → CP-BP-5**: `bug-fix-list.json` passes validation script with zero errors.
348
+ **Gate → CP-BP-5**: The draft batch validates and its merge dry run returns zero errors; the final list is still untouched.
352
349
 
353
- **Gate → CP-BP-6**: Local generated-plan review loaded `${SKILL_DIR}/references/generated-plan-review.md`, reviewed only newly added/changed bug entries, applied accepted fixes through draft/source data, regenerated, revalidated, and completed terminal draft cleanup or reported why recovery state was retained.
350
+ **Gate → CP-BP-6**: Local generated-plan review loaded `${SKILL_DIR}/references/generated-plan-review.md`, reviewed only the draft batch, applied accepted fixes through the draft, revalidated, returned `PASS`, then ran exactly one insert whose `appended_ids` match the batch and removed the draft.
354
351
 
355
352
  ## Local Generated-Plan Review Gate
356
353
 
357
- Run this gate after `.prizmkit/plans/bug-fix-list.json` passes generation and validation, and before the final handoff. It applies whenever this session added or changed final bug entries. For validate-only, summary-only, or draft-save flows with no final content change, report that the gate is not applicable.
354
+ Run this gate after the batch is written to the draft and passes draft validation, and before anything is written to `.prizmkit/plans/bug-fix-list.json`. The review target is the batch that is about to be inserted, not a file that already contains it. It applies whenever this session produced or changed a batch. For validate-only, summary-only, or draft-save flows with no batch, report that the gate is not applicable.
358
355
 
359
- Read `${SKILL_DIR}/references/generated-plan-review.md` and follow it as the single source of truth. Review the actual additions-only draft and use bounded `context` plus explicit-ID `show` for new or changed final entries because `.prizmkit` files may be gitignored; do not load unchanged historical entries. Repeat the Headless Context Completeness Gate for every reviewed new or changed bug. Keep the review planning-only, apply accepted findings to the draft/source representation, update ordinary appended entries through `append --replace-tail`, rerun validation, and repeat local review until it returns `PASS`. Complete §Draft Lifecycle cleanup before reporting the final written result.
356
+ Read `${SKILL_DIR}/references/generated-plan-review.md` and follow it as the single source of truth. Review the draft directly, reading `.prizmkit` artifacts from disk rather than from `git status` or `git diff`, because those files may be gitignored. Use bounded `context` and explicit-ID `show` only for already-committed entries whose contract or dependency semantics a reviewed item relies on; do not load unchanged historical entries. Repeat the Headless Context Completeness Gate for every reviewed new or changed bug. Keep the review planning-only: apply accepted findings to the draft, re-run draft validation and the merge dry run, and repeat local review until it returns `PASS`. Only then insert once and complete §Draft Lifecycle cleanup before reporting the final written result. If the insert is rejected after `PASS`, nothing landed; fix the draft, re-run validation and the dry run, and re-review only the entries you changed.
360
357
 
361
358
  #### Success Output
362
359
 
363
- Report the final path and bug counts, Local Generated-Plan Review verdict and reviewed IDs, accepted fixes, final validation result, draft cleanup result, and `bugfix-pipeline-launcher` handoff.
360
+ Report the final path and bug counts, Local Generated-Plan Review verdict and reviewed batch IDs, accepted fixes, the single insert's `appended_ids` and validation result, draft cleanup result, and `bugfix-pipeline-launcher` handoff.
364
361
 
365
362
  ### Checkpoints (Mandatory Gates)
366
363
 
@@ -372,8 +369,8 @@ Checkpoints catch cascading errors early — skipping one means the next phase b
372
369
  | **CP-BP-2** | Bugs Collected | All bugs extracted and confirmed by user | 2 |
373
370
  | **CP-BP-3** | Priorities Set | Severity/priority assigned, duplicates resolved | 3 |
374
371
  | **CP-BP-4** | Completeness Passed | All bugs pass headless execution readiness check | 4 |
375
- | **CP-BP-5** | File Generated | `bug-fix-list.json` passes validation script | 5 |
376
- | **CP-BP-6** | Local Plan Review Passed | Local generated-plan review loaded `${SKILL_DIR}/references/generated-plan-review.md`, reviewed only newly added/changed bug entries, applied accepted fixes through draft/source data, regenerated, revalidated, and completed terminal draft cleanup or reported why recovery state was retained | 5 |
372
+ | **CP-BP-5** | Draft Batch Ready | Additions-only or complete draft holds the whole batch with `context` IDs; merge dry run clean; final list still untouched | 5 |
373
+ | **CP-BP-6** | Reviewed, Then Inserted Once | Local generated-plan review loaded `${SKILL_DIR}/references/generated-plan-review.md`, reviewed only the draft batch, applied accepted fixes through the draft, revalidated, returned `PASS`, then ran exactly one insert whose `appended_ids` match the batch, and removed the draft | 5 |
377
374
 
378
375
  **Resume Detection**: If existing `.prizmkit/plans/bug-fix-list.json` or draft found, read `${SKILL_DIR}/references/error-recovery.md` §Resume Support for checkpoint-based resumption.
379
376
 
@@ -394,10 +391,10 @@ Batch-parse error logs with minimal interaction: parse automatically, resolve co
394
391
  - error_source: populated from log content
395
392
  - verification_type: default to `automated`
396
393
  - acceptance_criteria: auto-generate "Error no longer occurs in [scenario]"
397
- 4. Write `.prizmkit/plans/bug-fix-list.draft.json` as internal transport/recovery state.
394
+ 4. Write `.prizmkit/plans/bug-fix-list.draft.json` as the private review workspace.
398
395
  5. Run the Phase 4 Headless Context Completeness Gate on every parsed bug. Ask targeted clarification for missing expected behavior, selected scope, severity rationale, or verification decisions; update the draft until all entries pass.
399
- 6. Immediately use the Phase 5 writer: `generate` for a new complete list, or the bounded additions-only `append` command when a final list already exists. Do not add another broad approval after the consequential gaps are resolved.
400
- 7. If validation passes, run the mandatory Local Generated-Plan Review Gate, complete §Draft Lifecycle cleanup, and present the final written result. For later adjustments, recreate the minimum source draft from confirmed in-memory entries or exact-ID `show`, then rerun `generate` for a new-list rewrite or `append --replace-tail` for an ordinary append before recommending `bugfix-pipeline-launcher`.
396
+ 6. Check merge feasibility without writing: the complete draft for a new list, or `append --dry-run` when a final list already exists. Do not add another broad approval after the consequential gaps are resolved.
397
+ 7. Run the mandatory Local Generated-Plan Review Gate on the draft; on `PASS`, insert once with the matching Phase 5 writer, complete §Draft Lifecycle cleanup, and present the final written result. For later adjustments, recreate the minimum source draft from confirmed in-memory entries or exact-ID `show`, then rerun `generate` for a new-list rewrite or `append --replace-tail` for an ordinary append before recommending `bugfix-pipeline-launcher`.
401
398
 
402
399
  ## Operation: From Tests
403
400
 
@@ -409,10 +406,10 @@ Batch-parse failed test output:
409
406
  2. Parse each failed test case as a separate bug entry
410
407
  3. Auto-populate `failed_test_path`, `error_message`
411
408
  4. Set verification_type to `automated` (test already exists)
412
- 5. Write `.prizmkit/plans/bug-fix-list.draft.json` as internal transport/recovery state.
409
+ 5. Write `.prizmkit/plans/bug-fix-list.draft.json` as the private review workspace.
413
410
  6. Run the Phase 4 Headless Context Completeness Gate on every parsed failed-test bug; ask only for consequential missing expected behavior or scope, and allow needed edits, merges, or removals until all entries pass.
414
- 7. Immediately use the Phase 5 writer: `generate` for a new complete list, or the bounded additions-only `append` command when a final list already exists. Do not add another broad approval after the consequential gaps are resolved.
415
- 8. If validation passes, run the mandatory Local Generated-Plan Review Gate, complete §Draft Lifecycle cleanup, and present the final written result. For later adjustments, recreate the minimum source draft from confirmed in-memory entries or exact-ID `show`, then rerun `generate` for a new-list rewrite or `append --replace-tail` for an ordinary append before recommending `bugfix-pipeline-launcher`.
411
+ 7. Check merge feasibility without writing: the complete draft for a new list, or `append --dry-run` when a final list already exists. Do not add another broad approval after the consequential gaps are resolved.
412
+ 8. Run the mandatory Local Generated-Plan Review Gate on the draft; on `PASS`, insert once with the matching Phase 5 writer, complete §Draft Lifecycle cleanup, and present the final written result. For later adjustments, recreate the minimum source draft from confirmed in-memory entries or exact-ID `show`, then rerun `generate` for a new-list rewrite or `append --replace-tail` for an ordinary append before recommending `bugfix-pipeline-launcher`.
416
413
 
417
414
  ## Operation: Validate
418
415
 
@@ -462,12 +459,13 @@ Key points:
462
459
 
463
460
  Recommend invoking `bugfix-pipeline-launcher` to configure and launch the pipeline. Do NOT recommend running shell scripts directly — that is the launcher's responsibility.
464
461
 
465
- After `.prizmkit/plans/bug-fix-list.json` is generated, validated, and reviewed, present:
462
+ After `.prizmkit/plans/bug-fix-list.json` is inserted, present:
466
463
  1. Summary of generated bugs (count, severity breakdown)
467
- 2. Local generated-plan review result: verdict, reviewed bug IDs plus reviewed root prose when applicable, accepted fixes (or "none"), and final validation result
468
- 3. Draft cleanup result: `removed`, or retained path and recovery reason
469
- 4. Recommend: "Say 'start fixing' to launch the bugfix pipeline via `bugfix-pipeline-launcher`"
470
- 5. Alternative: fix a single bug interactively via `bug-fix-workflow`
464
+ 2. Local generated-plan review result: verdict, reviewed batch IDs plus reviewed root prose when applicable, and accepted fixes (or "none")
465
+ 3. Insert result: the single writer run and the `appended_ids` it reported
466
+ 4. Draft cleanup result: `removed`, or retained path and recovery reason
467
+ 5. Recommend: "Say 'start fixing' to launch the bugfix pipeline via `bugfix-pipeline-launcher`"
468
+ 6. Alternative: fix a single bug interactively via `bug-fix-workflow`
471
469
 
472
470
  ## Error Handling
473
471
 
@@ -1,28 +1,28 @@
1
1
  # Generated Plan Review — Bug Planner
2
2
 
3
- Use this local review loop after `.prizmkit/plans/bug-fix-list.json` has been generated and validated. This is a planner-output review, not an implementation/code-diff review.
3
+ Use this local review loop on the batch that is about to be inserted into `.prizmkit/plans/bug-fix-list.json`. This is a planner-output review, not an implementation/code-diff review, and it runs before the insert so the verdict decides whether the batch lands.
4
4
 
5
5
  ## Trigger
6
6
 
7
- Run only when the current `bug-planner` session generated, appended, or rewrote final bug planning content. For validate-only, summary-only, or draft-save flows with no new final planning content, report: `Local generated-plan review: not applicable — no new final planning content written.`
7
+ Run only when the current `bug-planner` session produced a draft batch of new or changed bugs. For validate-only, summary-only, or draft-save flows with no new batch, report: `Local generated-plan review: not applicable — no new planning content written.`
8
8
 
9
- In ordinary incremental flow, the additions draft is internal transport/recovery state. After consequential decisions and the pre-generation gate pass, immediately run `append`; do not ask for separate user approval merely to promote the draft into the final list. Keep the draft through the bounded review/fix loop, then delete the ordinary internal draft before final handoff only after this review returns `PASS` and final validation passes.
9
+ The draft is the whole review surface. Nothing is written to the final list while this loop runs: the single insert happens only after a `PASS`, so the verdict decides whether the batch lands and no final list ever holds reviewed-looking entries that were never reviewed.
10
10
 
11
11
  ## Direct Artifact Reads
12
12
 
13
13
  Planning artifacts may be gitignored, untracked, or absent from staged and unstaged diffs. Read the actual files through bounded, source-first access:
14
14
 
15
- - Final list: `.prizmkit/plans/bug-fix-list.json` — run the `context` command for root prose and recent IDs; do not load the complete historical list into AI context
16
- - Draft/source list when present: `.prizmkit/plans/bug-fix-list.draft.json` — ordinary incremental flow contains only additions and may be read completely
15
+ - Review target: `.prizmkit/plans/bug-fix-list.draft.json` — the additions-only or complete batch, read in full; this is the only planning content under review
16
+ - Final list: `.prizmkit/plans/bug-fix-list.json` — run the `context` command for root prose and recent IDs; do not load the complete historical list into AI context, and do not expect this session's batch to appear in it yet
17
17
  - Pre-session final list snapshot if captured in memory before edits — retain only bounded root context, prior task count/next ID, and the five recent `{id, description}` pairs, not full history
18
18
  - In-memory draft object if the session has not written a draft file yet
19
- - Newly appended or explicitly modified final entries — load each known ID with `python3 ${SKILL_DIR}/scripts/validate-bug-list.py show --input .prizmkit/plans/bug-fix-list.json --id B-NNN`
19
+ - Already-committed entries a reviewed item relies on — load each referenced ID with `python3 ${SKILL_DIR}/scripts/validate-bug-list.py show --input .prizmkit/plans/bug-fix-list.json --id B-NNN`
20
20
 
21
- Do not rely on `git status`, `git diff`, or `git diff --cached` to decide whether planner output exists or changed. Do not invoke `/prizmkit-code-review` for this planner-output gate; `/prizmkit-code-review` remains the implementation/code-diff review gate after `/prizmkit-implement`.
21
+ Do not rely on `git status`, `git diff`, or `git diff --cached` to decide whether planner output exists or changed. If the final list already contains this session's batch, the insert ran before the verdict — report that instead of reviewing an already-landed batch. Do not invoke `/prizmkit-code-review` for this planner-output gate; `/prizmkit-code-review` remains the implementation/code-diff review gate after `/prizmkit-implement`.
22
22
 
23
23
  ## Review Scope
24
24
 
25
- 1. For ordinary append flow, use the additions-only draft and the append result's stable bug IDs (`B-*`) as the change set; do not compare complete pre/post historical arrays in AI context.
25
+ 1. Treat the draft batch as the change set. Its `B-*` IDs are the sequential `context` IDs the writer demands at insert time, so the IDs reviewed are the IDs that land; do not compare complete pre/post historical arrays in AI context.
26
26
  2. Root project prose is reviewable content, not generator metadata. Read it through `context`; review new or changed `project_description` and `global_context` even when no bug item changed. Ignore root generator metadata and formatting-only differences such as `$schema`, `created_at`, and `created_by`.
27
27
  3. Treat bugs as changed only when item fields changed; load and review newly added bug IDs and explicitly changed bug IDs through `show`.
28
28
  4. Preserve unchanged historical bugs. Do not load unchanged entries merely to establish review completeness. Use the full-list validator for dependency/DAG truth; use `show` for an old dependency or overlap only when a specific reviewed bug requires semantic inspection.
@@ -48,6 +48,7 @@ For every reviewed bug entry, check:
48
48
  - Incident scope: keep one bug per task unless the same root cause or same failing contract is confirmed; do not domain-merge by module
49
49
  - Injected-field fidelity: confirmed actual/expected behavior and completion evidence are in `description` and `acceptance_criteria`; AI rewrite is allowed; dropping a confirmed item is not
50
50
  - Size: there is no word-count upper limit and no acceptance-criteria count upper limit
51
+ - Insert feasibility: every `dependencies` entry on a committed ID exists in the final list, the batch uses the consecutive IDs `context` reported, and the merge dry run returns `"valid": true`. Draft-only validation cannot see the merged graph, so a `PASS` without this check is a verdict the writer may reject.
51
52
  - Browser interaction and verification fields for UI-reproducible bugs when applicable
52
53
 
53
54
  ## Headless Context Completeness Gate
@@ -80,32 +81,27 @@ Do not add new JSON fields, context objects, schema versions, or pipeline runtim
80
81
  1. Present findings with IDs, severity, evidence, and suggested draft/source changes.
81
82
  2. Accept only findings that improve planner correctness, completeness, or downstream bug-fix execution readiness.
82
83
  3. Apply accepted fixes to `.prizmkit/plans/bug-fix-list.draft.json` or the in-memory draft object first.
83
- 4. For an ordinary incremental additions-only draft, rerun the existing script's append command:
84
+ 4. Re-check the fixed draft without writing:
84
85
 
85
86
  ```bash
86
- python3 ${SKILL_DIR}/scripts/validate-bug-list.py append --input .prizmkit/plans/bug-fix-list.draft.json --output .prizmkit/plans/bug-fix-list.json --replace-tail
87
+ python3 ${SKILL_DIR}/scripts/validate-bug-list.py append --input .prizmkit/plans/bug-fix-list.draft.json --output .prizmkit/plans/bug-fix-list.json --dry-run
87
88
  ```
88
89
 
89
- `--replace-tail` is allowed only for the exact final-list suffix whose IDs match this session's additions, so accepted local-review fixes replace the current session's entries without duplicating or loading history. If the suffix does not match exactly, stop rather than rewriting another task.
90
+ The dry run enforces the same base, exact-ID, dependency, and complete-DAG checks as the real insert, so keep it green instead of discovering a merge conflict after `PASS`. A new list or explicit full-list rewrite has no merged graph to check and nothing to overwrite, so its single post-`PASS` `generate` is the validator; if it rejects the draft, fix and retry, because no file was written.
90
91
 
91
- For a new list or explicit full-list rewrite, regenerate through the existing script's complete-draft command:
92
-
93
- ```bash
94
- python3 ${SKILL_DIR}/scripts/validate-bug-list.py generate --input .prizmkit/plans/bug-fix-list.draft.json --output .prizmkit/plans/bug-fix-list.json
95
- ```
96
-
97
- 5. Re-run append/generation until valid. Do not hand-patch `.prizmkit/plans/bug-fix-list.json` as the source of truth and do not add a user approval round for the mechanical rewrite.
92
+ 5. Never hand-patch `.prizmkit/plans/bug-fix-list.json` as the source of truth and do not add a user approval round for the mechanical rewrite.
98
93
  6. Repeat the local review and Headless Context Completeness Gate only for entries changed by accepted fixes.
99
- 7. Do not report `PASS` until every required fresh-session answer is available and final validation passes after regeneration.
100
- 8. After `PASS`, delete `.prizmkit/plans/bug-fix-list.draft.json` before final handoff when it is an ordinary internal draft. Retain it only for explicit draft-save, interruption or blocking before the terminal gates, or cleanup failure; report the retained path and reason. Cleanup failure leaves the reviewed final list valid but must be reported as stale recovery state.
94
+ 7. Do not report `PASS` until every required fresh-session answer is available and the draft validation plus merge dry run pass.
95
+ 8. After `PASS`, insert the batch exactly once (`append`, or `generate` for a new list), confirm the writer's `appended_ids`, then delete `.prizmkit/plans/bug-fix-list.draft.json`. Retain a draft only under the conditions in `${SKILL_DIR}/references/error-recovery.md`; report the retained path and reason.
101
96
 
102
97
  ## Report Format
103
98
 
104
99
  Report in the final handoff summary:
105
100
 
106
101
  - `Local generated-plan review: PASS` or `NEEDS_FIXES`
107
- - Reviewed bug IDs and reviewed root prose (`project_description` / `global_context`) when applicable
102
+ - Reviewed batch IDs and reviewed root prose (`project_description` / `global_context`) when applicable
103
+ - Merge dry-run result
108
104
  - Accepted fixes, or `none`
109
- - Final validation result
105
+ - Insert result: the single writer run and the `appended_ids` it reported
110
106
  - Draft cleanup result: `removed`, or retained path and recovery reason
111
- - Note when review was not applicable because no new final planning content was written
107
+ - Note when review was not applicable because no new planning content was written
@@ -15,6 +15,7 @@ Usage:
15
15
  python3 validate-bug-list.py context --input .prizmkit/plans/bug-fix-list.json
16
16
  python3 validate-bug-list.py show --input .prizmkit/plans/bug-fix-list.json --id B-001
17
17
  python3 validate-bug-list.py append --input additions.json --output .prizmkit/plans/bug-fix-list.json
18
+ python3 validate-bug-list.py append --input additions.json --output .prizmkit/plans/bug-fix-list.json --dry-run
18
19
  python3 validate-bug-list.py generate --input draft.json --output .prizmkit/plans/bug-fix-list.json
19
20
 
20
21
  Exit codes:
@@ -477,6 +478,11 @@ def cmd_append(args):
477
478
  _err(" " + error)
478
479
  return 1
479
480
 
481
+ if getattr(args, "dry_run", False):
482
+ _info("Dry run: {} bug(s) would be appended to {}".format(
483
+ len(prepared), args.output))
484
+ return 0
485
+
480
486
  try:
481
487
  _write_json_atomic(args.output, merged)
482
488
  except Exception as exc:
@@ -605,6 +611,11 @@ def main():
605
611
  action="store_true",
606
612
  help="Replace the exact existing suffix with the same IDs after local review",
607
613
  )
614
+ p_append.add_argument(
615
+ "--dry-run",
616
+ action="store_true",
617
+ help="Validate the would-be merged list without writing it",
618
+ )
608
619
 
609
620
  # -- generate --
610
621
  p_generate = subparsers.add_parser(
@@ -59,7 +59,7 @@ The user chose this skill intentionally. Respect that choice.
59
59
 
60
60
  ## Draft Lifecycle (Hard Rule)
61
61
 
62
- In an ordinary successful flow, `.prizmkit/plans/feature-list.draft.json` is internal writer/review input, not a user approval artifact or separate deliverable. Keep it through `generate`/`append`, validation, and the complete Local Generated-Plan Review fix loop. After the Local Generated-Plan Review returns `PASS` and final validation passes, delete the ordinary internal draft before final handoff. Report `Draft cleanup: removed`. Present the final written result. Retain the draft only for explicit draft-save, interruption or blocking before terminal gates, or cleanup failure; report its path and reason. Cleanup failure does not invalidate an already reviewed and validated final list, but makes the file stale recovery state. Never delete explicit recovery state merely because an older valid final exists. For later adjustments, recreate only the minimum source draft from confirmed in-memory entries or exact-ID `show`; ordinary appends must preserve the exact current suffix and use `append --replace-tail` without loading unrelated history.
62
+ In an ordinary successful flow, `.prizmkit/plans/feature-list.draft.json` is a private reasoning workspace: never presented as a deliverable, never given its own approval round, never retained after the batch lands. Build the batch there, validate it there, and run the entire Local Generated-Plan Review against it; the final list is written exactly once, after that review returns `PASS`, by a single `append` (existing list) or `generate` (new list). Then delete the draft immediately and report `Draft cleanup: removed`. A batch lost before it lands is regenerated from the confirmed decisions rather than recovered. Conditions that must keep a draft — explicit draft-save, interruption, a writer-rejected insert, cleanup failure — are defined in `${SKILL_DIR}/references/error-recovery.md`; name any retained path in the terminal report. For later adjustments, recreate only the minimum source draft from confirmed in-memory entries or exact-ID `show`; modifying entries this session did not create is the one case that still uses `append --replace-tail` for the exact current suffix.
63
63
 
64
64
  ## Fresh-Session Task Contract (Hard Rule)
65
65
 
@@ -242,8 +242,8 @@ For ordinary incremental planning, NEVER read the complete existing `.prizmkit/p
242
242
  2. Inspect those pairs only for material continuation, duplication, conflict, a shared public behavior contract, or an execution prerequisite. Add no historical content by default. Chronology, broad module overlap, shared technology, reusable implementation ideas, or similar wording alone do not establish relevance.
243
243
  3. Incorporate only the minimum exact information whose omission would make scope, behavior, compatibility, acceptance criteria, or dependency ordering incorrect. Reject vague `related to F-123` text and whole-description copying; add an old ID to `dependencies` only when it is a real execution prerequisite.
244
244
  4. Only for an explicitly identified old ID or a validation error, run `python3 ${SKILL_DIR}/scripts/validate-and-generate.py show --input .prizmkit/plans/feature-list.json --id F-NNN`.
245
- 5. For append-only planning, write `.prizmkit/plans/feature-list.draft.json` with exactly one root field, `features`, containing non-empty unchanged-schema items. This bounded draft is internal transport/recovery state, not a user approval artifact. Required draft fields are `id`, `title`, `description`, `priority`, `dependencies`, and `acceptance_criteria`; `status` may be omitted and defaults to `pending`; existing optional fields remain allowed. Copy no root metadata or historical features.
246
- 6. After consequential decisions and the pre-generation gate pass, immediately run `append`: `python3 ${SKILL_DIR}/scripts/validate-and-generate.py append --input .prizmkit/plans/feature-list.draft.json --output .prizmkit/plans/feature-list.json`. Do not ask for separate user approval merely to promote the draft into the final list. The command validates the base, IDs, dependencies, and complete DAG before atomic replacement. Apply review fixes only to the bounded additions and rerun `append --replace-tail` for the exact current suffix, then follow §Draft Lifecycle before presenting the final written result. `generate` remains the initial-plan/full-rewrite path.
245
+ 5. For append-only planning, write `.prizmkit/plans/feature-list.draft.json` with exactly one root field, `features`, containing non-empty unchanged-schema items. This bounded draft is a private reasoning workspace and the review target, not a user approval artifact or a deliverable. Required draft fields are `id`, `title`, `description`, `priority`, `dependencies`, and `acceptance_criteria`; `status` may be omitted and defaults to `pending`; existing optional fields remain allowed. Copy no root metadata or historical features.
246
+ 6. After consequential decisions and the pre-generation gate pass, write the batch to the additions-only draft with the `context` IDs and check merge feasibility without writing: `python3 ${SKILL_DIR}/scripts/validate-and-generate.py append --input .prizmkit/plans/feature-list.draft.json --output .prizmkit/plans/feature-list.json --dry-run`. Do not ask for separate user approval merely to promote the draft into the final list. The dry run applies the same base, ID, dependency, and DAG checks the real insert will, so a `PASS` review is a promise the insert keeps. Run the real `append` once, after §Local Generated-Plan Review Gate returns `PASS`, then follow §Draft Lifecycle. `generate` remains the initial-plan/full-rewrite path.
247
247
  ## Scenario Routing
248
248
  Classify user intent first:
249
249
 
@@ -311,11 +311,10 @@ Execute the planning workflow in conversation mode with mandatory checkpoints:
311
311
  4.2 **Browser interaction** (mandatory for fullstack/frontend projects) — see §Browser Interaction Planning below. Qualifying features get `browser_interaction` by default. Only confirm with the user as a batch summary; do NOT ask per-feature.
312
312
  5. Verify DAG/order/priorities and show the mandatory priority/complexity calibration table with short rationales for every item
313
313
  6. Run the pre-generation Headless Context Completeness Gate (see §Pre-Generation Completeness Review below)
314
- 7. Build or append `.prizmkit/plans/feature-list.json`
314
+ 7. Build the batch in `.prizmkit/plans/feature-list.draft.json` with the `context` IDs; write no final list yet
315
315
  8. Apply default testing strategy (see §Testing Defaults below)
316
- 9. Validate and fix until pass
317
- 10. Run the mandatory Local Generated-Plan Review Gate (see §Local Generated-Plan Review Gate below); apply accepted fixes to the draft/source plan, regenerate, and revalidate until pass, then complete §Draft Lifecycle cleanup
318
- 11. Summarize final feature table with local review result, validation outcome, and draft cleanup result
316
+ 9. Validate the draft batch and fix until pass, then run the mandatory Local Generated-Plan Review Gate (see §Local Generated-Plan Review Gate below) against it; apply accepted fixes to the draft, revalidate, and repeat until `PASS`
317
+ 11. Insert the batch once (`append`, or `generate` for a new list), complete §Draft Lifecycle cleanup, and summarize the final feature table with the inserted IDs, local review verdict, writer result, and cleanup result
319
318
 
320
319
  ### Checkpoints (Mandatory Gates)
321
320
 
@@ -329,9 +328,9 @@ Checkpoints catch cascading errors early — skipping one means the next phase b
329
328
  | **CP-FP-3.1** | Browser Interaction Applied | Qualifying features have `browser_interaction` field; user confirmed or opted out | 4 |
330
329
  | **CP-FP-3.2** | Testing Defaults Applied | All features have appropriate testing expectations | 7 |
331
330
  | **CP-FP-3.3** | Completeness Review Passed | Every new or changed feature passes fresh-session simulation from `description` and `acceptance_criteria` without conversation history or `user_context` | 6 |
332
- | **CP-FP-4** | `.prizmkit/plans/feature-list.json` Generated | Schema validates, all required keys present | 7-8 |
333
- | **CP-FP-5** | Final Validation Pass | Python script returns `"valid": true` with zero errors | 9 |
334
- | **CP-FP-6** | Local Plan Review Passed | Local generated-plan review loaded `${SKILL_DIR}/references/generated-plan-review.md`, reviewed only newly added/changed feature entries, applied accepted fixes through draft/source data, regenerated, revalidated, and completed terminal draft cleanup or reported why recovery state was retained | 10-11 |
331
+ | **CP-FP-4** | Draft Batch Ready | Additions-only or complete draft holds the whole batch with `context` IDs; final list still untouched | 7-8 |
332
+ | **CP-FP-5** | Draft Validation Pass | Draft validation plus the merge dry run return `"valid": true` with zero errors | 9 |
333
+ | **CP-FP-6** | Reviewed, Then Inserted Once | Local generated-plan review loaded `${SKILL_DIR}/references/generated-plan-review.md`, reviewed only the draft batch, applied accepted fixes through the draft, revalidated, returned `PASS`, then ran exactly one `append`/`generate` whose `appended_ids` match the batch, and removed the draft | 10-11 |
335
334
 
336
335
  **Resume Detection**: If existing artifacts are found, read `${SKILL_DIR}/references/error-recovery.md` §Resume Support for checkpoint-based resumption.
337
336
 
@@ -378,12 +377,12 @@ For simple incremental planning, skip detailed Phase 2-3 analysis:
378
377
  3. Generate next sequential feature IDs
379
378
  4. Draft features (title + description + acceptance_criteria + dependencies)
380
379
  5. Run the mandatory pre-generation Headless Context Completeness Gate from Phase 6 on every drafted feature; repair the draft until each returns `PASS`.
381
- 6. Write the additions-only draft as internal input and immediately call the append command without a second draft-approval pause:
380
+ 6. Write the additions-only draft as internal input — no second draft-approval pause — and check merge feasibility without writing:
382
381
  ```bash
383
- python3 ${SKILL_DIR}/scripts/validate-and-generate.py append --input .prizmkit/plans/feature-list.draft.json --output .prizmkit/plans/feature-list.json
382
+ python3 ${SKILL_DIR}/scripts/validate-and-generate.py append --input .prizmkit/plans/feature-list.draft.json --output .prizmkit/plans/feature-list.json --dry-run
384
383
  ```
385
- 7. If valid → run the mandatory Local Generated-Plan Review Gate, complete §Draft Lifecycle cleanup, then present the final written result and recommend the next step
386
- 8. If invalid → apply fixes to the additions-only draft, re-run `append` (max 2 attempts, then escalate to full workflow)
384
+ 7. If valid → run the mandatory Local Generated-Plan Review Gate on the draft, insert once with `append`, complete §Draft Lifecycle cleanup, then present the final written result and recommend the next step
385
+ 8. If invalid → apply fixes to the additions-only draft and re-run the dry run (max 2 attempts, then escalate to full workflow)
387
386
 
388
387
  ## Browser Interaction Planning
389
388
 
@@ -433,13 +432,13 @@ Key requirements:
433
432
  python3 ${SKILL_DIR}/scripts/validate-and-generate.py append --input .prizmkit/plans/feature-list.draft.json --output .prizmkit/plans/feature-list.json
434
433
  ```
435
434
 
436
- `generate` fills root defaults for complete drafts. `append` preserves root metadata and historical features, defaults missing new statuses, validates the merged full list, and writes only after validation passes. In ordinary incremental flow, call `append` immediately after semantic gates; the draft is not a second approval checkpoint. If either command fails, fix its source draft and retry. Do not delete the draft until the Local Generated-Plan Review and final validation pass; then follow §Draft Lifecycle.
435
+ `generate` fills root defaults for complete drafts. `append` preserves root metadata and historical features, defaults missing new statuses, validates the merged full list, and writes only after validation passes. In ordinary incremental flow the draft is not a second approval checkpoint, but it is the review target, so the real `append` runs once, after the Local Generated-Plan Review returns `PASS`; use `append --dry-run` for every earlier merged check because it validates without writing. If the writer rejects the batch, fix its source draft and retry — a rejection never leaves a partial final list. Then follow §Draft Lifecycle.
437
436
 
438
437
  ## Local Generated-Plan Review Gate
439
438
 
440
- Run this gate after `.prizmkit/plans/feature-list.json` passes generation and validation, and before the final handoff. It applies whenever this session added or changed final feature entries. For validate-only, summary-only, or draft-save flows with no final content change, report that the gate is not applicable.
439
+ Run this gate after the batch is written to the draft and passes draft validation, and before anything is written to `.prizmkit/plans/feature-list.json`: the review target is the batch about to be inserted, not a file that already contains it. For validate-only, summary-only, or draft-save flows with no batch, report that the gate is not applicable.
441
440
 
442
- Read `${SKILL_DIR}/references/generated-plan-review.md` and follow it as the single source of truth. Review the actual additions-only draft and use bounded `context` plus explicit-ID `show` for new or changed final entries because `.prizmkit` files may be gitignored; do not load unchanged historical entries. Repeat the Headless Context Completeness Gate for every reviewed new or changed feature. Keep the review planning-only, apply accepted findings to the draft/source representation, update ordinary appended entries through `append --replace-tail`, rerun validation, and repeat the local review until it returns `PASS`. Complete §Draft Lifecycle cleanup before reporting the final written result.
441
+ Read `${SKILL_DIR}/references/generated-plan-review.md` and follow it as the single source of truth. Review the draft directly, reading `.prizmkit` artifacts from disk rather than from `git status` or `git diff`, because those files may be gitignored. Use bounded `context` and explicit-ID `show` only for already-committed entries whose contract or dependency semantics a reviewed item relies on; do not load unchanged historical entries. Repeat the Headless Context Completeness Gate for every reviewed new or changed feature. Keep the review planning-only: apply accepted findings to the draft, re-run draft validation and the merge dry run, and repeat until it returns `PASS`. Only then insert once, complete §Draft Lifecycle cleanup, and present the final written result. If the insert is rejected after `PASS`, nothing landed: fix the draft, re-run validation and the dry run, and re-review the changed entries.
443
442
 
444
443
  ## Testing Defaults (Phase 8)
445
444
 
@@ -485,11 +484,12 @@ When the session appears to be ending:
485
484
 
486
485
  ## Handoff Message Template
487
486
 
488
- After successful validation and the Local Generated-Plan Review Gate, report:
487
+ After the batch is inserted, report:
489
488
  1. Output file path
490
489
  2. Total features + newly added features
491
490
  3. Dependency and priority highlights
492
- 4. Local generated-plan review result: verdict, reviewed feature IDs plus reviewed root prose when applicable, accepted fixes (or "none"), final validation result, and draft cleanup result (`removed` or retained path/reason)
493
- 5. Recommended next action: `feature-pipeline-launcher`
491
+ 4. Local generated-plan review result: verdict, reviewed batch IDs plus reviewed root prose when applicable, and accepted fixes (or "none")
492
+ 5. Insert result: the writer's `appended_ids` from the single post-review run, plus its validation output and the draft cleanup result (`removed`, or retained path and reason)
493
+ 6. Recommended next action: `feature-pipeline-launcher`
494
494
 
495
495
  Then apply the post-planning handoff in §Scope Boundary (ask before proceeding; never auto-execute).
@@ -1,28 +1,28 @@
1
1
  # Generated Plan Review — Feature Planner
2
2
 
3
- Use this local review loop after `.prizmkit/plans/feature-list.json` has been generated and validated. This is a planner-output review, not an implementation/code-diff review.
3
+ Use this local review loop on the batch that is about to be inserted into `.prizmkit/plans/feature-list.json`. This is a planner-output review, not an implementation/code-diff review, and it runs before the insert so the verdict decides whether the batch lands.
4
4
 
5
5
  ## Trigger
6
6
 
7
- Run only when the current `feature-planner` session generated, appended, or rewrote final feature planning content. For validate-only, summary-only, or draft-save flows with no new final planning content, report: `Local generated-plan review: not applicable — no new final planning content written.`
7
+ Run only when the current `feature-planner` session produced a draft batch of new or changed features. For validate-only, summary-only, or draft-save flows with no new batch, report: `Local generated-plan review: not applicable — no new planning content written.`
8
8
 
9
- In ordinary incremental flow, the additions draft is internal transport/recovery state. After consequential decisions and the pre-generation gate pass, immediately run `append`; do not ask for separate user approval merely to promote the draft into the final list. Keep the draft through the bounded review/fix loop, then delete the ordinary internal draft before final handoff only after this review returns `PASS` and final validation passes.
9
+ The draft is the whole review surface. Nothing is written to the final list while this loop runs: the single insert happens only after a `PASS`, so the verdict decides whether the batch lands and no final list ever holds reviewed-looking entries that were never reviewed.
10
10
 
11
11
  ## Direct Artifact Reads
12
12
 
13
13
  Planning artifacts may be gitignored, untracked, or absent from staged and unstaged diffs. Read the actual files through bounded, source-first access:
14
14
 
15
- - Final list: `.prizmkit/plans/feature-list.json` — run the `context` command for root prose and recent IDs; do not load the complete historical list into AI context
16
- - Draft/source list when present: `.prizmkit/plans/feature-list.draft.json` — ordinary incremental flow contains only additions and may be read completely
15
+ - Review target: `.prizmkit/plans/feature-list.draft.json` — the additions-only or complete batch, read in full; this is the only planning content under review
16
+ - Final list: `.prizmkit/plans/feature-list.json` — run the `context` command for root prose and recent IDs; do not load the complete historical list into AI context, and do not expect this session's batch to appear in it yet
17
17
  - Pre-session final list snapshot if captured in memory before edits — retain only bounded root context, prior task count/next ID, and the five recent `{id, description}` pairs, not full history
18
18
  - In-memory draft object if the session has not written a draft file yet
19
- - Newly appended or explicitly modified final entries — load each known ID with `python3 ${SKILL_DIR}/scripts/validate-and-generate.py show --input .prizmkit/plans/feature-list.json --id F-NNN`
19
+ - Already-committed entries a reviewed item relies on — load each referenced ID with `python3 ${SKILL_DIR}/scripts/validate-and-generate.py show --input .prizmkit/plans/feature-list.json --id F-NNN`
20
20
 
21
- Do not rely on `git status`, `git diff`, or `git diff --cached` to decide whether planner output exists or changed. Do not invoke `/prizmkit-code-review` for this planner-output gate; `/prizmkit-code-review` remains the implementation/code-diff review gate after `/prizmkit-implement`.
21
+ Do not rely on `git status`, `git diff`, or `git diff --cached` to decide whether planner output exists or changed. If the final list already contains this session's batch, the insert ran before the verdict — report that instead of reviewing an already-landed batch. Do not invoke `/prizmkit-code-review` for this planner-output gate; `/prizmkit-code-review` remains the implementation/code-diff review gate after `/prizmkit-implement`.
22
22
 
23
23
  ## Review Scope
24
24
 
25
- 1. For ordinary append flow, use the additions-only draft and the append result's stable feature IDs (`F-*`) as the change set; do not compare complete pre/post historical arrays in AI context.
25
+ 1. Treat the draft batch as the change set. Its `F-*` IDs are the sequential `context` IDs the writer demands at insert time, so the IDs reviewed are the IDs that land; do not compare complete pre/post historical arrays in AI context.
26
26
  2. Root project prose is reviewable content, not generator metadata. Read it through `context`; review new or changed `project_description` and `global_context` even when no feature item changed. Ignore root generator metadata and formatting-only differences such as `$schema`, `created_at`, and `created_by`.
27
27
  3. Treat features as changed only when item fields changed; load and review newly added feature IDs and explicitly changed feature IDs through `show`.
28
28
  4. Preserve unchanged historical features. Do not load unchanged entries merely to establish review completeness. Use the full-list validator for dependency/DAG truth; use `show` for an old dependency only when a specific reviewed feature requires semantic inspection of that prerequisite.
@@ -49,6 +49,7 @@ For every reviewed feature entry, check:
49
49
  - Unused nested work: new or changed items MUST NOT store required work only in `user_context`; move it into `description` and `acceptance_criteria`
50
50
  - Injected-field fidelity: confirmed WHAT, constraints, non-goals, boundaries, and completion evidence are in `description` and `acceptance_criteria`; AI rewrite is allowed; dropping a confirmed item is not
51
51
  - Size: there is no word-count upper limit and no acceptance-criteria count upper limit; long contracts are not defects
52
+ - Insert feasibility: every `dependencies` entry on a committed ID exists in the final list, the batch uses the consecutive IDs `context` reported, and the merge dry run returns `"valid": true`. Draft-only validation cannot see the merged graph, so a `PASS` without this check is a verdict the writer may reject.
52
53
  - Browser interaction fields for qualifying frontend/fullstack features when applicable
53
54
 
54
55
  ## Headless Context Completeness Gate
@@ -82,32 +83,27 @@ Do not add new JSON fields, context objects, schema versions, or pipeline runtim
82
83
  1. Present findings with IDs, severity, evidence, and suggested draft/source changes.
83
84
  2. Accept only findings that improve planner correctness, completeness, or downstream execution readiness.
84
85
  3. Apply accepted fixes to `.prizmkit/plans/feature-list.draft.json` or the in-memory draft object first.
85
- 4. For an ordinary incremental additions-only draft, rerun the existing script's append command:
86
+ 4. Re-check the fixed draft without writing:
86
87
 
87
88
  ```bash
88
- python3 ${SKILL_DIR}/scripts/validate-and-generate.py append --input .prizmkit/plans/feature-list.draft.json --output .prizmkit/plans/feature-list.json --replace-tail
89
+ python3 ${SKILL_DIR}/scripts/validate-and-generate.py append --input .prizmkit/plans/feature-list.draft.json --output .prizmkit/plans/feature-list.json --dry-run
89
90
  ```
90
91
 
91
- `--replace-tail` is allowed only for the exact final-list suffix whose IDs match this session's additions, so accepted local-review fixes replace the current session's entries without duplicating or loading history. If the suffix does not match exactly, stop rather than rewriting another task.
92
+ The dry run enforces the same base, exact-ID, dependency, and complete-DAG checks as the real insert, so keep it green instead of discovering a merge conflict after `PASS`. An initial plan or explicit full-list rewrite has no merged graph to check and nothing to overwrite, so its single post-`PASS` `generate` is the validator; if it rejects the draft, fix and retry, because no file was written.
92
93
 
93
- For an initial plan or explicit full-list rewrite, regenerate through the existing script's complete-draft command:
94
-
95
- ```bash
96
- python3 ${SKILL_DIR}/scripts/validate-and-generate.py generate --input .prizmkit/plans/feature-list.draft.json --output .prizmkit/plans/feature-list.json
97
- ```
98
-
99
- 5. Re-run append/generation until valid. Do not hand-patch `.prizmkit/plans/feature-list.json` as the source of truth and do not add a user approval round for the mechanical rewrite.
94
+ 5. Never hand-patch `.prizmkit/plans/feature-list.json` as the source of truth and do not add a user approval round for the mechanical rewrite.
100
95
  6. Repeat the local review and Headless Context Completeness Gate only for entries changed by accepted fixes.
101
- 7. Do not report `PASS` until every required fresh-session answer is available and final validation passes after regeneration.
102
- 8. After `PASS`, delete `.prizmkit/plans/feature-list.draft.json` before final handoff when it is an ordinary internal draft. Retain it only for explicit draft-save, interruption or blocking before the terminal gates, or cleanup failure; report the retained path and reason. Cleanup failure leaves the reviewed final list valid but must be reported as stale recovery state.
96
+ 7. Do not report `PASS` until every required fresh-session answer is available and the draft validation plus merge dry run pass.
97
+ 8. After `PASS`, insert the batch exactly once (`append`, or `generate` for a new list), confirm the writer's `appended_ids`, then delete `.prizmkit/plans/feature-list.draft.json`. Retain a draft only under the conditions in `${SKILL_DIR}/references/error-recovery.md`; report the retained path and reason.
103
98
 
104
99
  ## Report Format
105
100
 
106
101
  Report in the final handoff summary:
107
102
 
108
103
  - `Local generated-plan review: PASS` or `NEEDS_FIXES`
109
- - Reviewed feature IDs and reviewed root prose (`project_description` / `global_context`) when applicable
104
+ - Reviewed batch IDs and reviewed root prose (`project_description` / `global_context`) when applicable
105
+ - Merge dry-run result
110
106
  - Accepted fixes, or `none`
111
- - Final validation result
107
+ - Insert result: the single writer run and the `appended_ids` it reported
112
108
  - Draft cleanup result: `removed`, or retained path and recovery reason
113
- - Note when review was not applicable because no new final planning content was written
109
+ - Note when review was not applicable because no new planning content was written
@@ -83,21 +83,21 @@ The additions-only `.prizmkit/plans/feature-list.draft.json` has exactly one roo
83
83
  }
84
84
  ```
85
85
 
86
- Each item uses the unchanged final feature-item schema. Required draft fields are `id`, `title`, `description`, `priority`, `dependencies`, and `acceptance_criteria`; `status` may be omitted and defaults to `pending`; all existing optional feature fields remain allowed. Root project metadata and historical features must not be copied into this draft. Treat it as internal transport/recovery state, not as a user approval artifact.
86
+ Each item uses the unchanged final feature-item schema. Required draft fields are `id`, `title`, `description`, `priority`, `dependencies`, and `acceptance_criteria`; `status` may be omitted and defaults to `pending`; all existing optional feature fields remain allowed. Root project metadata and historical features must not be copied into this draft. Treat it as a private reasoning workspace and the review target, not as a user approval artifact or a deliverable.
87
87
 
88
88
  ### Step 4: Rebalance Priority
89
89
 
90
90
  Default to append only. Reprioritizing historical tasks is an explicit old-task modification, not part of ordinary append flow; require the user to identify the affected IDs and load only those tasks with `show`.
91
91
 
92
- ### Step 5: Append and Validate
92
+ ### Step 5: Check Feasibility, Review, Then Insert Once
93
93
 
94
- After consequential decisions and the pre-generation semantic gate pass, immediately merge the additions-only draft with untouched history outside AI context:
94
+ After consequential decisions and the pre-generation semantic gate pass, check the additions against untouched history outside AI context without writing:
95
95
 
96
96
  ```bash
97
- python3 ${SKILL_DIR}/scripts/validate-and-generate.py append --input .prizmkit/plans/feature-list.draft.json --output .prizmkit/plans/feature-list.json
97
+ python3 ${SKILL_DIR}/scripts/validate-and-generate.py append --input .prizmkit/plans/feature-list.draft.json --output .prizmkit/plans/feature-list.json --dry-run
98
98
  ```
99
99
 
100
- The append command validates the existing base, exact sequential new IDs, all old/new dependencies, and the complete DAG before atomically replacing the final list. Do not ask for separate user approval merely to promote the additions draft into the final list. Fix invalid additions and rerun until valid; during local review, correct the exact current suffix with `append --replace-tail`. Keep the draft through that fix loop. After review returns `PASS` and final validation passes, delete the ordinary internal draft before final handoff. For a later adjustment, recreate only the exact current additions from confirmed in-memory entries or explicit-ID `show` results before using `append --replace-tail`. Use `generate` only for an initial plan or an explicit full-list rewrite.
100
+ The dry run checks the existing base, exact sequential new IDs, all old/new dependencies, and the complete DAG exactly as the real insert will. Do not ask for separate user approval merely to promote the additions draft into the final list, and do not insert yet: run the Local Generated-Plan Review against the draft, apply accepted fixes there, and re-run this dry run until it stays green and the review returns `PASS`. Then insert once with the real `append` and delete the draft. For a later adjustment, recreate only the exact current additions from confirmed in-memory entries or explicit-ID `show` results before using `append --replace-tail`. Use `generate` only for an initial plan or an explicit full-list rewrite.
101
101
 
102
102
  ## Merge/Rewrite Rules
103
103
 
@@ -125,6 +125,6 @@ Use concise prompts during interaction:
125
125
  - [ ] Additions-only draft contains no root metadata or historical features
126
126
  - [ ] Atomic append ran immediately after semantic gates without a post-draft approval pause
127
127
  - [ ] Existing project conventions preserved from project evidence
128
- - [ ] Complete merged dependency graph validates through `append`
129
- - [ ] Ordinary internal draft was deleted after review and final validation, or its retained recovery path and reason were reported
128
+ - [ ] Complete merged dependency graph validates through the `append` dry run
129
+ - [ ] Batch was reviewed in the draft, inserted exactly once after `PASS`, and the draft was deleted or its retained recovery path and reason were reported
130
130
  - [ ] Next step recommendation: `feature-pipeline-launcher`
@@ -54,14 +54,14 @@ Split into additional top-level features only when:
54
54
 
55
55
  ### Phase 6: Generate + Validate
56
56
 
57
- 1. Write `.prizmkit/plans/feature-list.draft.json` as internal writer/review input after the main Skill's consequential decision and pre-generation gates pass.
58
- 2. Immediately generate the final file without a separate draft-approval pause:
57
+ 1. Write `.prizmkit/plans/feature-list.draft.json` as the private review workspace after the main Skill's consequential decision and pre-generation gates pass.
58
+ 2. Run the main Skill's mandatory Local Generated-Plan Review against that complete draft and apply accepted fixes through the draft. Nothing is written to `feature-list.json` yet.
59
+ 3. After review returns `PASS`, generate the final file once without a separate draft-approval pause:
59
60
  ```bash
60
61
  python3 ${SKILL_DIR}/scripts/validate-and-generate.py generate --input .prizmkit/plans/feature-list.draft.json --output .prizmkit/plans/feature-list.json
61
62
  ```
62
- 3. Fix draft errors, then re-run `generate` until validation passes.
63
- 4. Run the main Skill's mandatory Local Generated-Plan Review and apply accepted fixes through the complete draft.
64
- 5. After review returns `PASS` and final validation passes, delete the ordinary internal draft before final handoff. Retain it only under the main Skill's explicit recovery conditions and report the reason.
63
+ 4. If `generate` rejects the draft, fix the draft and retry; a rejection writes no file, so the final list is never partial.
64
+ 5. After the single successful `generate`, delete the draft and report `Draft cleanup: removed`. Retain it only under the main Skill's explicit recovery conditions and report the reason.
65
65
 
66
66
  ## Quality Rules
67
67
 
@@ -80,6 +80,6 @@ Split into additional top-level features only when:
80
80
  - [ ] User confirmed initial feature scope
81
81
  - [ ] IDs are sequential
82
82
  - [ ] `status` initialized to `pending`
83
- - [ ] Validation and Local Generated-Plan Review pass
84
- - [ ] Ordinary internal draft was deleted, or its retained recovery path and reason were reported
83
+ - [ ] Validation and Local Generated-Plan Review passed on the draft, then exactly one `generate` ran
84
+ - [ ] Ordinary internal draft was deleted after that single insert, or its retained recovery path and reason were reported
85
85
  - [ ] Next step recommendation: `feature-pipeline-launcher`
@@ -19,6 +19,7 @@ Usage:
19
19
  python3 validate-and-generate.py context --input .prizmkit/plans/feature-list.json
20
20
  python3 validate-and-generate.py show --input .prizmkit/plans/feature-list.json --id F-001
21
21
  python3 validate-and-generate.py append --input additions.json --output .prizmkit/plans/feature-list.json
22
+ python3 validate-and-generate.py append --input additions.json --output .prizmkit/plans/feature-list.json --dry-run
22
23
  python3 validate-and-generate.py generate --input draft.json --output .prizmkit/plans/feature-list.json
23
24
  python3 validate-and-generate.py summary --input .prizmkit/plans/feature-list.json [--format markdown|json]
24
25
  python3 validate-and-generate.py grade --workspace /.codebuddy/skill-evals/feature-planner-workspace --iteration iteration-1
@@ -837,6 +838,11 @@ def cmd_append(args):
837
838
  _err(" " + error)
838
839
  return 1
839
840
 
841
+ if getattr(args, "dry_run", False):
842
+ _info("Dry run: {} feature(s) would be appended to {}".format(
843
+ len(prepared), args.output))
844
+ return 0
845
+
840
846
  try:
841
847
  _write_json_atomic(args.output, merged)
842
848
  except Exception as exc:
@@ -1113,6 +1119,11 @@ def main():
1113
1119
  action="store_true",
1114
1120
  help="Replace the exact existing suffix with the same IDs after local review",
1115
1121
  )
1122
+ p_append.add_argument(
1123
+ "--dry-run",
1124
+ action="store_true",
1125
+ help="Validate the would-be merged list without writing it",
1126
+ )
1116
1127
 
1117
1128
  # -- generate --
1118
1129
  p_generate = subparsers.add_parser(
@@ -56,7 +56,7 @@ The user chose this skill intentionally. Respect that choice.
56
56
 
57
57
  ## Draft Lifecycle (Hard Rule)
58
58
 
59
- In an ordinary successful flow, `.prizmkit/plans/refactor-list.draft.json` is internal writer/review input, not a user approval artifact or separate deliverable. Keep it through `generate`/`append`, validation, and the complete Local Generated-Plan Review fix loop. After the Local Generated-Plan Review returns `PASS` and final validation passes, delete the ordinary internal draft before final handoff. Report `Draft cleanup: removed`. Present the final written result only after cleanup succeeds or a retained recovery reason is surfaced. Retain the draft only for explicit draft-save, interruption or blocking before terminal gates, or cleanup failure; report its path and reason. Cleanup failure does not invalidate an already reviewed and validated final list, but makes the file stale recovery state. Never delete explicit recovery state merely because an older valid final exists. For later adjustments, recreate only the minimum source draft from confirmed in-memory entries or exact-ID `show`; ordinary appends must preserve the exact current suffix and use `append --replace-tail` without loading unrelated history.
59
+ In an ordinary successful flow, `.prizmkit/plans/refactor-list.draft.json` is a private reasoning workspace: never presented as a deliverable, never given its own approval round, never retained after the batch lands. Build the batch there, validate it there, and run the entire Local Generated-Plan Review against it; the final list is written exactly once, after that review returns `PASS`, by a single `append` (existing list) or `generate` (new list). Then delete the draft immediately and report `Draft cleanup: removed`. A batch lost before it lands is regenerated from the confirmed decisions rather than recovered. Conditions that must keep a draft — explicit draft-save, interruption, a writer-rejected insert, cleanup failure — are defined in `${SKILL_DIR}/references/error-recovery.md`, and any retained path is named in the terminal report. For later adjustments, recreate only the minimum source draft from confirmed in-memory entries or exact-ID `show`; modifying entries this session did not create is the one case that still uses `append --replace-tail` for the exact current suffix.
60
60
 
61
61
  ## Fresh-Session Task Contract (Hard Rule)
62
62
 
@@ -172,12 +172,12 @@ For any flow that appends to an existing `.prizmkit/plans/refactor-list.json`, N
172
172
  ```bash
173
173
  python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py show --input .prizmkit/plans/refactor-list.json --id R-NNN
174
174
  ```
175
- 5. For append-only planning, write `.prizmkit/plans/refactor-list.draft.json` as an additions-only object with exactly one root field, `refactors`. The array must be non-empty; each item uses the unchanged final refactor-item schema, with required `id`, `title`, `description`, `scope`, `type`, `priority`, `complexity`, `behavior_preservation`, `dependencies`, and `acceptance_criteria`; `status` may be omitted and defaults to `pending`; all existing optional refactor fields remain allowed. Do not copy root metadata or historical refactors into this bounded internal transport/recovery state.
176
- 6. After consequential decisions and the pre-generation gate pass, immediately run `append` outside AI context:
175
+ 5. For append-only planning, write `.prizmkit/plans/refactor-list.draft.json` as an additions-only object with exactly one root field, `refactors`. The array must be non-empty; each item uses the unchanged final refactor-item schema, with required `id`, `title`, `description`, `scope`, `type`, `priority`, `complexity`, `behavior_preservation`, `dependencies`, and `acceptance_criteria`; `status` may be omitted and defaults to `pending`; all existing optional refactor fields remain allowed. Do not copy root metadata or historical refactors into this bounded private reasoning workspace, which is the review target rather than a deliverable.
176
+ 6. After consequential decisions and the pre-generation gate pass, write the batch to the additions-only draft with the `context` IDs and check merge feasibility without writing:
177
177
  ```bash
178
- python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py append --input .prizmkit/plans/refactor-list.draft.json --output .prizmkit/plans/refactor-list.json
178
+ python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py append --input .prizmkit/plans/refactor-list.draft.json --output .prizmkit/plans/refactor-list.json --dry-run
179
179
  ```
180
- Do not ask for separate user approval merely to promote the draft into the final list. The command validates the base, IDs, dependencies, and complete DAG before atomic replacement. Apply review fixes only to the bounded additions and rerun `append --replace-tail` for the exact current suffix, then follow §Draft Lifecycle before presenting the final written result. Existing `generate` remains the new-list/full-rewrite path.
180
+ Do not ask for separate user approval merely to promote the draft into the final list. The dry run applies the same base, exact-ID, dependency, and complete-DAG checks the real insert will, so a `PASS` review is a promise the insert keeps. Run the real `append` once, after §Local Generated-Plan Review Gate returns `PASS`, then follow §Draft Lifecycle. Existing `generate` remains the new-list/full-rewrite path.
181
181
 
182
182
  ## Operation Modes
183
183
 
@@ -319,25 +319,22 @@ See `${SKILL_DIR}/references/planning-phases.md` for the full 5-step review chec
319
319
  **Goal**: Produce `.prizmkit/plans/refactor-list.json` and validate it.
320
320
 
321
321
  **IMPORTANT: Do NOT hand-write the final JSON file.** Instead:
322
- 1. For a new list, write the complete `.prizmkit/plans/refactor-list.draft.json`. For an existing list, follow §Bounded Existing-List Context and write only the additions-only `refactors` object as internal transport/recovery state.
323
- 2. Once required semantic gates pass, immediately use the matching deterministic writer without a second draft-approval pause:
324
- - New list or explicit full-list rewrite:
325
- ```bash
326
- python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py generate --input .prizmkit/plans/refactor-list.draft.json --output .prizmkit/plans/refactor-list.json
327
- ```
322
+ 1. For a new list, write the complete `.prizmkit/plans/refactor-list.draft.json`. For an existing list, follow §Bounded Existing-List Context and write only the additions-only `refactors` object. The draft is the review target, so nothing is written to the final list yet.
323
+ 2. Once required semantic gates pass, check merge feasibility without a second draft-approval pause:
324
+ - New list or explicit full-list rewrite: the draft is complete and nothing exists to overwrite, so its post-review `generate` is the validator.
328
325
  - Ordinary existing-list append:
329
326
  ```bash
330
- python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py append --input .prizmkit/plans/refactor-list.draft.json --output .prizmkit/plans/refactor-list.json
327
+ python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py append --input .prizmkit/plans/refactor-list.draft.json --output .prizmkit/plans/refactor-list.json --dry-run
331
328
  ```
332
- `generate` fills root defaults for a complete draft. `append` preserves root metadata and historical refactors, defaults missing new statuses, and writes only after the merged full list validates.
329
+ The dry run applies the same base, exact-ID, dependency, and complete-DAG checks the real insert will, including `append`'s preservation of root metadata and historical refactors.
333
330
  3. If validation fails -> fix the draft and retry (max 3 attempts)
334
- 4. If validation passes -> run the mandatory Local Generated-Plan Review Gate (see §Local Generated-Plan Review Gate below); apply accepted fixes to the draft/source plan, use `append --replace-tail` for ordinary appended entries, and revalidate until pass
335
- 5. If review and validation pass -> delete the ordinary internal draft under §Draft Lifecycle; retain and report it only for an explicit recovery condition or cleanup failure
331
+ 4. If the draft and dry run pass -> run the mandatory Local Generated-Plan Review Gate (see §Local Generated-Plan Review Gate below) on the draft; apply accepted fixes to the draft and revalidate until `PASS`
332
+ 5. After `PASS` -> insert exactly once (`append`, or `generate` for a new list), confirm its `appended_ids`, and delete the draft under §Draft Lifecycle
336
333
  6. Present the final written result, summary, and draft cleanup result. For later user adjustments, recreate the minimum source draft from confirmed in-memory entries or exact-ID `show`, then rerun `generate` for a new-list rewrite or `append --replace-tail` for ordinary appended entries
337
334
 
338
- **CHECKPOINT CP-RP-6**: `.prizmkit/plans/refactor-list.json` generated and validated.
335
+ **CHECKPOINT CP-RP-6**: The draft batch validates and its merge dry run returns zero errors; the final list is still untouched.
339
336
 
340
- **CHECKPOINT CP-RP-7**: Local generated-plan review loaded `${SKILL_DIR}/references/generated-plan-review.md`, reviewed only newly added/changed refactor entries, applied accepted fixes through draft/source data, regenerated, revalidated, and completed terminal draft cleanup or reported why recovery state was retained.
337
+ **CHECKPOINT CP-RP-7**: Local generated-plan review loaded `${SKILL_DIR}/references/generated-plan-review.md`, reviewed only the draft batch, applied accepted fixes through the draft, revalidated, returned `PASS`, then ran exactly one insert whose `appended_ids` match the batch and removed the draft.
341
338
 
342
339
  ## Checkpoints (Mandatory Gates)
343
340
 
@@ -349,14 +346,14 @@ See `${SKILL_DIR}/references/planning-phases.md` for the full 5-step review chec
349
346
  | **CP-RP-3** | Items Decomposed | All items have deps, complexity, preservation strategy | 4 |
350
347
  | **CP-RP-4** | Items Confirmed | User confirmed/modified/skipped each item | 5 |
351
348
  | **CP-RP-5** | Completeness OK | DAG valid, preservation strategies declared, no gaps | 6 |
352
- | **CP-RP-6** | Output Valid | `.prizmkit/plans/refactor-list.json` passes validation script | 7 |
353
- | **CP-RP-7** | Local Plan Review Passed | Local generated-plan review loaded `${SKILL_DIR}/references/generated-plan-review.md`, reviewed only newly added/changed refactor entries, applied accepted fixes through draft/source data, regenerated, revalidated, and completed terminal draft cleanup or reported why recovery state was retained | 7 |
349
+ | **CP-RP-6** | Draft Batch Ready | Additions-only or complete draft holds the whole batch with `context` IDs; merge dry run clean; final list still untouched | 7 |
350
+ | **CP-RP-7** | Reviewed, Then Inserted Once | Local generated-plan review loaded `${SKILL_DIR}/references/generated-plan-review.md`, reviewed only the draft batch, applied accepted fixes through the draft, revalidated, returned `PASS`, then ran exactly one insert whose `appended_ids` match the batch, and removed the draft | 7 |
354
351
 
355
352
  ## Local Generated-Plan Review Gate
356
353
 
357
- Run this gate after `.prizmkit/plans/refactor-list.json` passes generation and validation, and before the final handoff. It applies whenever this session added or changed final refactor entries. For validate-only, summary-only, or draft-save flows with no final content change, report that the gate is not applicable.
354
+ Run this gate after the batch is written to the draft and passes draft validation, and before anything is written to `.prizmkit/plans/refactor-list.json`. The review target is the batch that is about to be inserted, not a file that already contains it. It applies whenever this session produced or changed a batch. For validate-only, summary-only, or draft-save flows with no batch, report that the gate is not applicable.
358
355
 
359
- Read `${SKILL_DIR}/references/generated-plan-review.md` and follow it as the single source of truth. Review the actual additions-only draft and use bounded `context` plus explicit-ID `show` for new or changed final entries because `.prizmkit` files may be gitignored; do not load unchanged historical entries. Repeat the Headless Context Completeness Gate for every reviewed new or changed refactor. Keep the review planning-only, apply accepted findings to the draft/source representation, update ordinary appended entries through `append --replace-tail`, rerun validation, and repeat local review until it returns `PASS`. Complete §Draft Lifecycle cleanup before reporting the final written result.
356
+ Read `${SKILL_DIR}/references/generated-plan-review.md` and follow it as the single source of truth. Review the draft directly, reading `.prizmkit` artifacts from disk rather than from `git status` or `git diff`, because those files may be gitignored. Use bounded `context` and explicit-ID `show` only for already-committed entries whose contract or dependency semantics a reviewed item relies on; do not load unchanged historical entries. Repeat the Headless Context Completeness Gate for every reviewed new or changed refactor. Keep the review planning-only: apply accepted findings to the draft, re-run draft validation and the merge dry run, and repeat local review until it returns `PASS`. Only then insert once and complete §Draft Lifecycle cleanup before reporting the final written result. If the insert is rejected after `PASS`, nothing landed; fix the draft, re-run validation and the dry run, and re-review only the entries you changed.
360
357
 
361
358
  ## Output Rules
362
359
 
@@ -459,11 +456,12 @@ When validation fails, existing planning artifacts are detected, the session is
459
456
 
460
457
  ## Handoff Message Template
461
458
 
462
- After successful validation and the Local Generated-Plan Review Gate, report:
459
+ After the batch is inserted, report:
463
460
  1. Output file path
464
461
  2. Total refactor items
465
462
  3. Dependency ordering highlights (which items must run first)
466
463
  4. Behavior preservation strategy distribution (N items with test-gate, M with snapshot, etc.)
467
- 5. Local generated-plan review result: verdict, reviewed refactor IDs plus reviewed root prose when applicable, accepted fixes (or "none"), and final validation result
468
- 6. Draft cleanup result: `removed`, or retained path and recovery reason
469
- 7. Recommended next action: `refactor-pipeline-launcher`
464
+ 5. Local generated-plan review result: verdict, reviewed batch IDs plus reviewed root prose when applicable, and accepted fixes (or "none")
465
+ 6. Insert result: the single writer run and the `appended_ids` it reported
466
+ 7. Draft cleanup result: `removed`
467
+ 8. Recommended next action: `refactor-pipeline-launcher`
@@ -31,13 +31,13 @@ For simple refactoring with minimal scope.
31
31
  **NEVER proceed without explicit user selection via `AskUserQuestion`. Do NOT render options as plain text — the user must be able to click/select.**
32
32
  3. Draft items (title + type + scope + description + acceptance_criteria + behavior_preservation + dependencies)
33
33
  4. Run the mandatory Phase 6 Headless Context Completeness Gate on every drafted item; repair the draft until each returns `PASS` without conversation history or `user_context`.
34
- 5. Write `.prizmkit/plans/refactor-list.draft.json` as internal transport/recovery state. For a new list, use the complete final-list draft and `generate`; for an existing list, write only the additions-only `refactors` object and immediately call `append` without a second draft-approval pause:
34
+ 5. Write `.prizmkit/plans/refactor-list.draft.json` as the private review workspace. For a new list, write the complete final-list draft; for an existing list, write only the additions-only `refactors` object and check merge feasibility without writing, with no second draft-approval pause:
35
35
  ```bash
36
- python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py append --input .prizmkit/plans/refactor-list.draft.json --output .prizmkit/plans/refactor-list.json
36
+ python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py append --input .prizmkit/plans/refactor-list.draft.json --output .prizmkit/plans/refactor-list.json --dry-run
37
37
  ```
38
- 6. If valid -> run the mandatory Local Generated-Plan Review Gate from the main `refactor-planner` skill. After review returns `PASS` and final validation passes, delete the ordinary internal draft under the main Skill's Draft Lifecycle, then present the final written result before recommending `refactor-pipeline-launcher`.
38
+ 6. If valid -> run the mandatory Local Generated-Plan Review Gate from the main `refactor-planner` skill against the draft. After review returns `PASS`, insert once (`append`, or `generate` for a new list), delete the draft under the main Skill's Draft Lifecycle, then present the final written result before recommending `refactor-pipeline-launcher`.
39
39
  7. For a later adjustment after cleanup, recreate only the minimum source draft from confirmed in-memory entries or exact-ID `show`, then rerun `generate` for a new-list rewrite or `append --replace-tail` for an ordinary append.
40
- 8. If invalid -> apply fixes to the draft and re-run the same writer (`generate` for a new list, `append` for an existing list; max 2 attempts, then escalate to full workflow)
40
+ 8. If invalid -> apply fixes to the draft and re-run the dry run; a new list is validated by its post-`PASS` `generate`, which writes nothing when it rejects (max 2 attempts, then escalate to full workflow)
41
41
 
42
42
  ## When NOT to Use Fast Path
43
43
  - More than 2 refactor items
@@ -1,28 +1,28 @@
1
1
  # Generated Plan Review — Refactor Planner
2
2
 
3
- Use this local review loop after `.prizmkit/plans/refactor-list.json` has been generated and validated. This is a planner-output review, not an implementation/code-diff review.
3
+ Use this local review loop on the batch that is about to be inserted into `.prizmkit/plans/refactor-list.json`. This is a planner-output review, not an implementation/code-diff review, and it runs before the insert so the verdict decides whether the batch lands.
4
4
 
5
5
  ## Trigger
6
6
 
7
- Run only when the current `refactor-planner` session generated, appended, or rewrote final refactor planning content. For validate-only, summary-only, or draft-save flows with no new final planning content, report: `Local generated-plan review: not applicable — no new final planning content written.`
7
+ Run only when the current `refactor-planner` session produced a draft batch of new or changed refactors. For validate-only, summary-only, or draft-save flows with no new batch, report: `Local generated-plan review: not applicable — no new planning content written.`
8
8
 
9
- In ordinary incremental flow, the additions draft is internal transport/recovery state. After consequential decisions and the pre-generation gate pass, immediately run `append`; do not ask for separate user approval merely to promote the draft into the final list. Keep the draft through the bounded review/fix loop, then delete the ordinary internal draft before final handoff only after this review returns `PASS` and final validation passes.
9
+ The draft is the whole review surface. Nothing is written to the final list while this loop runs: the single insert happens only after a `PASS`, so the verdict decides whether the batch lands and no final list ever holds reviewed-looking entries that were never reviewed.
10
10
 
11
11
  ## Direct Artifact Reads
12
12
 
13
13
  Planning artifacts may be gitignored, untracked, or absent from staged and unstaged diffs. Read the actual files through bounded, source-first access:
14
14
 
15
- - Final list: `.prizmkit/plans/refactor-list.json` — run the `context` command for root prose and recent IDs; do not load the complete historical list into AI context
16
- - Draft/source list when present: `.prizmkit/plans/refactor-list.draft.json` — ordinary incremental flow contains only additions and may be read completely
15
+ - Review target: `.prizmkit/plans/refactor-list.draft.json` — the additions-only or complete batch, read in full; this is the only planning content under review
16
+ - Final list: `.prizmkit/plans/refactor-list.json` — run the `context` command for root prose and recent IDs; do not load the complete historical list into AI context, and do not expect this session's batch to appear in it yet
17
17
  - Pre-session final list snapshot if captured in memory before edits — retain only bounded root context, prior task count/next ID, and the five recent `{id, description}` pairs, not full history
18
18
  - In-memory draft object if the session has not written a draft file yet
19
- - Newly appended or explicitly modified final entries — load each known ID with `python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py show --input .prizmkit/plans/refactor-list.json --id R-NNN`
19
+ - Already-committed entries a reviewed item relies on — load each referenced ID with `python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py show --input .prizmkit/plans/refactor-list.json --id R-NNN`
20
20
 
21
- Do not rely on `git status`, `git diff`, or `git diff --cached` to decide whether planner output exists or changed. Do not invoke `/prizmkit-code-review` for this planner-output gate; `/prizmkit-code-review` remains the implementation/code-diff review gate after `/prizmkit-implement`.
21
+ Do not rely on `git status`, `git diff`, or `git diff --cached` to decide whether planner output exists or changed. If the final list already contains this session's batch, the insert ran before the verdict — report that instead of reviewing an already-landed batch. Do not invoke `/prizmkit-code-review` for this planner-output gate; `/prizmkit-code-review` remains the implementation/code-diff review gate after `/prizmkit-implement`.
22
22
 
23
23
  ## Review Scope
24
24
 
25
- 1. For ordinary append flow, use the additions-only draft and the append result's stable refactor IDs (`R-*`) as the change set; do not compare complete pre/post historical arrays in AI context.
25
+ 1. Treat the draft batch as the change set. Its `R-*` IDs are the sequential `context` IDs the writer demands at insert time, so the IDs reviewed are the IDs that land; do not compare complete pre/post historical arrays in AI context.
26
26
  2. Root project prose is reviewable content, not generator metadata. Read it through `context`; review new or changed `project_description` and `global_context` even when no refactor item changed. Ignore root generator metadata and formatting-only differences such as `$schema`, `created_at`, and `created_by`.
27
27
  3. Treat refactors as changed only when item fields changed; load and review newly added refactor IDs and explicitly changed refactor IDs through `show`.
28
28
  4. Preserve unchanged historical refactors. Do not load unchanged entries merely to establish review completeness. Use the full-list validator for dependency/DAG truth; use `show` for an old dependency or behavior-preservation interaction only when a specific reviewed refactor requires semantic inspection.
@@ -48,6 +48,7 @@ For every reviewed refactor entry, check:
48
48
  - Priority/complexity calibration: every reviewed refactor has short priority and complexity rationales, medium/low assignments are allowed when criteria fit, and high/high is not used merely because the work is cleanup, framework-related, multi-file, or user-requested
49
49
  - Domain cohesion: same contract-bounded capability stays one top-level item; independently shippable is not a split reason; same-domain over-split is `NEEDS_FIXES` unless one existing headless session cannot finish
50
50
  - Injected-field fidelity: confirmed structure, scope, preservation, and completion evidence are in `description` and `acceptance_criteria`; AI rewrite is allowed; dropping a confirmed item is not
51
+ - Insert feasibility: every `dependencies` entry on a committed ID exists in the final list, the batch uses the consecutive IDs `context` reported, and the merge dry run returns `"valid": true`. Draft-only validation cannot see the merged graph, so a `PASS` without this check is a verdict the writer may reject.
51
52
  - Size: there is no word-count upper limit and no acceptance-criteria count upper limit
52
53
 
53
54
  ## Headless Context Completeness Gate
@@ -80,32 +81,27 @@ Do not add new JSON fields, context objects, schema versions, or pipeline runtim
80
81
  1. Present findings with IDs, severity, evidence, and suggested draft/source changes.
81
82
  2. Accept only findings that improve planner correctness, completeness, or downstream behavior-preserving execution readiness.
82
83
  3. Apply accepted fixes to `.prizmkit/plans/refactor-list.draft.json` or the in-memory draft object first.
83
- 4. For an ordinary incremental additions-only draft, rerun the existing script's append command:
84
+ 4. Re-check the fixed draft without writing:
84
85
 
85
86
  ```bash
86
- python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py append --input .prizmkit/plans/refactor-list.draft.json --output .prizmkit/plans/refactor-list.json --replace-tail
87
+ python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py append --input .prizmkit/plans/refactor-list.draft.json --output .prizmkit/plans/refactor-list.json --dry-run
87
88
  ```
88
89
 
89
- `--replace-tail` is allowed only for the exact final-list suffix whose IDs match this session's additions, so accepted local-review fixes replace the current session's entries without duplicating or loading history. If the suffix does not match exactly, stop rather than rewriting another task.
90
+ The dry run enforces the same base, exact-ID, dependency, and complete-DAG checks as the real insert, so keep it green instead of discovering a merge conflict after `PASS`. A new list or explicit full-list rewrite has no merged graph to check and nothing to overwrite, so its single post-`PASS` `generate` is the validator; if it rejects the draft, fix and retry, because no file was written.
90
91
 
91
- For a new list or explicit full-list rewrite, regenerate through the existing script's complete-draft command:
92
-
93
- ```bash
94
- python3 ${SKILL_DIR}/scripts/validate-and-generate-refactor.py generate --input .prizmkit/plans/refactor-list.draft.json --output .prizmkit/plans/refactor-list.json
95
- ```
96
-
97
- 5. Re-run append/generation until valid. Do not hand-patch `.prizmkit/plans/refactor-list.json` as the source of truth and do not add a user approval round for the mechanical rewrite.
92
+ 5. Never hand-patch `.prizmkit/plans/refactor-list.json` as the source of truth and do not add a user approval round for the mechanical rewrite.
98
93
  6. Repeat the local review and Headless Context Completeness Gate only for entries changed by accepted fixes.
99
- 7. Do not report `PASS` until every required fresh-session answer is available and final validation passes after regeneration.
100
- 8. After `PASS`, delete `.prizmkit/plans/refactor-list.draft.json` before final handoff when it is an ordinary internal draft. Retain it only for explicit draft-save, interruption or blocking before the terminal gates, or cleanup failure; report the retained path and reason. Cleanup failure leaves the reviewed final list valid but must be reported as stale recovery state.
94
+ 7. Do not report `PASS` until every required fresh-session answer is available and the draft validation plus merge dry run pass.
95
+ 8. After `PASS`, insert the batch exactly once (`append`, or `generate` for a new list), confirm the writer's `appended_ids`, then delete `.prizmkit/plans/refactor-list.draft.json`. Retain a draft only under the conditions in `${SKILL_DIR}/references/error-recovery.md`; report the retained path and reason.
101
96
 
102
97
  ## Report Format
103
98
 
104
99
  Report in the final handoff summary:
105
100
 
106
101
  - `Local generated-plan review: PASS` or `NEEDS_FIXES`
107
- - Reviewed refactor IDs and reviewed root prose (`project_description` / `global_context`) when applicable
102
+ - Reviewed batch IDs and reviewed root prose (`project_description` / `global_context`) when applicable
103
+ - Merge dry-run result
108
104
  - Accepted fixes, or `none`
109
- - Final validation result
105
+ - Insert result: the single writer run and the `appended_ids` it reported
110
106
  - Draft cleanup result: `removed`, or retained path and recovery reason
111
- - Note when review was not applicable because no new final planning content was written
107
+ - Note when review was not applicable because no new planning content was written
@@ -18,6 +18,7 @@ Usage:
18
18
  python3 validate-and-generate-refactor.py context --input .prizmkit/plans/refactor-list.json
19
19
  python3 validate-and-generate-refactor.py show --input .prizmkit/plans/refactor-list.json --id R-001
20
20
  python3 validate-and-generate-refactor.py append --input additions.json --output .prizmkit/plans/refactor-list.json
21
+ python3 validate-and-generate-refactor.py append --input additions.json --output .prizmkit/plans/refactor-list.json --dry-run
21
22
  python3 validate-and-generate-refactor.py generate --input draft.json --output .prizmkit/plans/refactor-list.json
22
23
  python3 validate-and-generate-refactor.py summary --input .prizmkit/plans/refactor-list.json [--format markdown|json]
23
24
 
@@ -888,6 +889,11 @@ def cmd_append(args):
888
889
  _err(" " + error)
889
890
  return 1
890
891
 
892
+ if getattr(args, "dry_run", False):
893
+ _info("Dry run: {} refactor(s) would be appended to {}".format(
894
+ len(prepared), args.output))
895
+ return 0
896
+
891
897
  try:
892
898
  _write_json_atomic(args.output, merged)
893
899
  except Exception as exc:
@@ -1053,6 +1059,11 @@ def main():
1053
1059
  action="store_true",
1054
1060
  help="Replace the exact existing suffix with the same IDs after local review",
1055
1061
  )
1062
+ p_append.add_argument(
1063
+ "--dry-run",
1064
+ action="store_true",
1065
+ help="Validate the would-be merged list without writing it",
1066
+ )
1056
1067
 
1057
1068
  # -- generate --
1058
1069
  p_generate = subparsers.add_parser(
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "prizmkit",
3
- "version": "1.1.189",
3
+ "version": "1.1.190",
4
4
  "description": "Create a new PrizmKit-powered project with clean initialization — no framework dev files, just what you need.",
5
5
  "type": "module",
6
6
  "bin": {