@dzhechkov/keysarium 1.5.29 β†’ 1.5.30

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.
package/README.md CHANGED
@@ -29,6 +29,7 @@ After installation, open Claude Code in your project directory and start using s
29
29
  | Component | Count | Description |
30
30
  |-----------|-------|-------------|
31
31
  | **Skills** | 12 | `explore`, `frontend-design`, `goap-research-ed25519`, `presentation-storyteller`, `problem-solver-enhanced`, `reverse-engineering-unicorn`, `knowledge-extractor`, `ai-factory-mapper`, `analyst-manual-full`, `edu-site-generator`, `transcript-site-generator`, `feature-adr` |
32
+ | | | *bundled `feature-adr` at pipeline v1.3.60: amendment-Confirmation discipline (AM-N β†’ named falsifying test), derived `🚦 Gates` checkpoint line, opt-in Step-10 Delivery Gate (4-plane post-implementation review)* |
32
33
  | **Commands** | 20 | `/casarium`, `/new-research`, `/parallel-research`, `/discovery`, `/explore-case`, `/research`, `/cjm-prototype`, `/solve`, `/architecture-phase`, `/presentation`, `/harvest`, `/brain-export`, `/brain-import`, `/bto-benchmark`, `/feature-adr`, `/init-platform`, `/learning-stats`, `/dream`, `/workers`, `/analyst-manual` |
33
34
  | **Rules** | 14 | `research-quality`, `checkpoint-protocol`, `agent-swarm`, `domain-specific`, `anti-patterns`, `file-conventions`, `modular-reuse`, `feedback-loops`, `model-routing`, `trust-tiers`, `background-workers`, `dream-cycles`, `reward-learning`, `feature-adr-conventions` |
34
35
  | **Shards** | 9 | Context shards for each phase: `phase-0-discovery`, `phase-1-explore`, `phase-2-research`, `phase-25-cjm`, `phase-3-solve`, `phase-4-architecture`, `phase-5-presentation`, `phase-ai-factory`, `feature-adr` |
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@dzhechkov/keysarium",
3
- "version": "1.5.29",
3
+ "version": "1.5.30",
4
4
  "description": "AI Case Research Toolkit for Claude Code β€” full 7-phase pipeline with feature-adr, dream cycles, background workers, multi-platform support, reward learning, and Cloud.ru AI Factory mapping (--ai-factory flag)",
5
5
  "main": "src/cli.js",
6
6
  "bin": {
@@ -53,7 +53,8 @@ Trigger on:
53
53
  β”‚ β”œβ”€β”€ 06-implementation-plan.md ← Step 6: SPARC-GOAP planning (enhanced)
54
54
  β”‚ β”œβ”€β”€ 07-code.md ← Step 7: Code generation
55
55
  β”‚ β”œβ”€β”€ 08-qe.md ← Step 8: QE + brutal-honesty review (enhanced)
56
- β”‚ └── 09-fleet-qe.md ← Step 9: Fleet QE assessment (NEW)
56
+ β”‚ β”œβ”€β”€ 09-fleet-qe.md ← Step 9: Fleet QE assessment (NEW)
57
+ β”‚ └── 10-delivery-gate.md ← Step 10: Delivery Gate (opt-in, post-implementation)
57
58
  β”œβ”€β”€ references/
58
59
  β”‚ β”œβ”€β”€ complexity-matrix.md ← S/M/L/XL criteria & step activation
59
60
  β”‚ β”œβ”€β”€ adr-template.md ← ADR document template
@@ -140,6 +141,7 @@ Step | S | M | L | XL | Model | Agentic QE Skill
140
141
  7 Code | βœ“ | βœ“ | βœ“ | βœ“ | opus | β€” |
141
142
  8 QE + Brutal Honesty | βœ“ | βœ“ | βœ“ | βœ“ | sonnet | brutal-honesty-review |
142
143
  9 Fleet QE Assessment | - | - | βœ“ | βœ“ | sonnet | qe-req-val + risk-based + integration + regression + coverage |
144
+ 10 Delivery Gate (OPT-IN) | o | o | o | o | cross-family of coder | 4 planes: regressions β€– security β€– code-quality β€– product-honesty (o = runs only when explicitly requested; absent β‡’ byte-identical) |
143
145
  ```
144
146
 
145
147
  ## DAG Dependencies
@@ -271,6 +273,7 @@ features/<feature-slug>/
271
273
  | Step 7 | `<promise>FEATURE_ADR_IMPLEMENTED</promise>` |
272
274
  | Step 8 | `<promise>FEATURE_ADR_VERIFIED</promise>` |
273
275
  | Step 9 | `<promise>FEATURE_ADR_FLEET_VERIFIED</promise>` |
276
+ | Step 10 (opt-in) | `<promise>FEATURE_ADR_DELIVERY_GATED</promise>` |
274
277
 
275
278
  ## External Skills (loaded as needed)
276
279
 
@@ -589,7 +592,7 @@ npx @dzhechkov/skills-feature-adr init --with-learning --knowledge-extractor
589
592
  <promise>[PROMISE_TAG]</promise>
590
593
  Tier: {COMPLEXITY_TIER} | Active Steps: {ACTIVE_STEPS}
591
594
  πŸŽ“ Learning: {recalled} patterns recalled for this feature, {stored} new stored this run
592
- 🚦 Gates: challenge-panel βœ“ Β· claim-check βœ“ Β· discrimination not-run Β· amendments βœ“ Β· fleet β€”
595
+ 🚦 Gates: challenge-panel βœ“ Β· claim-check βœ“ Β· discrimination not-run Β· amendments βœ“ Β· fleet β€” Β· delivery n/a
593
596
 
594
597
  [2-3 line summary]
595
598
  Artifacts: [list] βœ…
@@ -0,0 +1,63 @@
1
+ # Step 10: Delivery Gate (opt-in) β€” review the feature as a PUBLISHED ENTITY
2
+
3
+ > **Opt-in.** This step runs only when explicitly requested (the user asks for a delivery gate, or the
4
+ > workflow form receives `args.deliveryGate: true` / `args.models.delivery`). Absent β‡’ the pipeline is
5
+ > byte-identical to a run without Step 10 β€” no agents, no artifact.
6
+
7
+ ## Why a separate step (not more Step-8)
8
+
9
+ Step 8 reviews the code as the coder's counterpart β€” pre-hand-off, working-tree view. Step 10 reviews the
10
+ feature as a **published entity**: the landed diff, its docs and claims, its behavior as a consumer will
11
+ meet it. Its distinctive plane β€” **product honesty** β€” is absent from Step 8 entirely: fabricated
12
+ completeness (output presented as complete when a source was unavailable), a feature that does less than
13
+ its description, misleading user-facing text. Origin: a real incident where a 2-agent fresh-eyes pass said
14
+ SHIP and only the full multi-plane review found the fabricated-completeness and cap-violation defects.
15
+
16
+ ## Protocol
17
+
18
+ ### 1. Four orthogonal planes (parallel, cross-family of the coder)
19
+
20
+ Each plane reads `07_code_changes/change_manifest.md` + the actual changed files (+ `git diff` for anything
21
+ uncommitted), calibrating on `architecture/vision.md` / `architecture/degradations.md` **when present** (an
22
+ accepted degradation is NOT a finding) and staying generic when absent:
23
+
24
+ | Plane | Hunts |
25
+ |---|---|
26
+ | **Regressions** | broken consumers/contracts; NEW I/O on previously-pure startup/lifespan/health paths without a negative resource-down test; removed/weakened tests (fixture-swap); silent semantic changes to shared surfaces |
27
+ | **Security** | injection via interpolated paths/refs/commands; secrets in code/artifacts; path traversal/symlink escapes; fail-open where the contract says fail-closed |
28
+ | **Code quality** | god-object growth; parallel implementations vs the reuse map; structurally-dead safeguards; swallowed errors |
29
+ | **Product honesty + common sense** | claims not backed by behavior; fabricated completeness; docs/READMEs promising more than the code does; misleading degradation/limits text |
30
+
31
+ Findings shape: `{severity: BLOCKER|HIGH|MED|LOW, title, where (file:line), why}` β€” confirmed only.
32
+
33
+ ### 2. Cross-validation (BLOCKER/HIGH, by index)
34
+
35
+ Every BLOCKER/HIGH is independently cross-validated by a second agent, matched **positionally by index β€”
36
+ never by title** (duplicate titles cross-contaminate). Default-to-FP when uncertain. A validator outage
37
+ surfaces the findings **UNVALIDATED** β€” never silently dropped, never silently confirmed.
38
+
39
+ ### 3. Machine-checkable hand-off criterion
40
+
41
+ Write `features/<slug>/10_delivery_review.md`:
42
+ - **## Verdict** β€” `hand-off: ready | blocked` (ready ⇔ 0 BLOCKER and 0 unwaived HIGH, cross-validation complete)
43
+ - **## Findings** β€” table: severity | plane | title | where | why | crossValidated
44
+ - **## Hand-off criterion** β€” machine-checkable rows: `0 BLOCKER: PASS/FAIL(n)`, `0 unwaived HIGH: PASS/FAIL(n)`.
45
+ Projects with a merge-request flow add THEIR OWN rows here (`CI terminal: …`, `draftβ†’ready: …`) β€” these are
46
+ **document rows the owner fills**, never API calls made by the pipeline. No merge-request flow β‡’ the rows
47
+ stay `β€”`.
48
+ - **## Note** β€” ADVISORY: the owner decides; nothing auto-aborts.
49
+
50
+ ### Hard rules
51
+
52
+ 1. **Findings only.** Step 10 NEVER posts to a VCS host, tracker, or any external service. Publishing a
53
+ review anywhere is a separate, explicit user instruction.
54
+ 2. **Advisory.** `hand-off: blocked` is a report, not an abort (a false gate kills trust).
55
+ 3. **Cross-family.** The planes run on the other model family than the coder (the Step-8 rule) β€” a model
56
+ must not deliver-gate its own code.
57
+ 4. **Cost control.** One gate per hand-off, not per commit. A stack of related changes gets one Step-10 pass
58
+ on the combined diff.
59
+
60
+ ## Promise tag
61
+
62
+ `<promise>FEATURE_ADR_DELIVERY_GATED</promise>` β€” and the `🚦 Gates:` line gains `delivery ready|blocked`
63
+ (derived from the review doc's Verdict, `n/a` when the gate was not requested).