cclaw-cli 0.5.12 → 0.5.13

package/dist/content/examples.js

@@ -311,106 +311,55 @@ Execution rule: complete and verify each wave before starting the next wave.
 | AC-2 (idempotency) | T-1, T-2 |
 | AC-3 (failure visibility) | T-3 |
 
+### Risk Assessment
+
+| Task/Wave | Risk | Likelihood | Impact | Mitigation |
+| --- | --- | --- | --- | --- |
+| T-3 (Wave 3) | SSE reconnect logic complex | Medium | High | Spike reconnect in isolation before integrating with feed UI |
+| Wave 2 → 3 | Publisher API contract may shift | Low | Medium | Pin contract in T-1 schema; T-2 integration test validates |
+
 ### WAIT_FOR_CONFIRM
 - Status: pending
 - Confirmed by:`,
-    tdd: `### RED test (Vitest) — written before production code
-
-\`\`\`typescript
-import { describe, it, expect } from "vitest";
-import { summarizeDedupedFeed } from "../notificationFeed";
-
-describe("summarizeDedupedFeed", () => {
-  it("counts unique keys and unread items", () => {
-    const summary = summarizeDedupedFeed([
-      { dedupeKey: "a", read: false },
-      { dedupeKey: "a", read: true },
-      { dedupeKey: "b", read: false },
-    ]);
-
-    expect(summary).toEqual({ uniqueKeys: 2, unread: 1 });
-  });
-});
-\`\`\`
-
-### Expected output (FAIL)
-
-\`\`\`bash
- FAIL  src/notificationFeed.test.ts
-Error: Cannot find module '../notificationFeed' imported from src/notificationFeed.test.ts
-\`\`\`
-
-> **Annotation:** This test MUST fail before any production code is written.
-
-### Iron Law verification
-
-1. **Run** the test command (for example: \`pnpm vitest run src/notificationFeed.test.ts\`).
-2. **Read output** and confirm the failure is due to the module/function not existing (or the function throwing “not implemented”), not due to a typo in assertions.
-3. **Confirm** the failure reason matches the intentional gap: **missing implementation**, not a flaky environment or misconfigured test runner.
-
-### Common mistakes to avoid
-
-- “GREEN” that secretly imports a helper that already implements the behavior (that is skipping RED).
-- Assertions that pass because the function returns \`undefined\` and the matcher is too loose.
-
-### GREEN (minimal implementation to pass RED)
-
-\`\`\`typescript
-export type FeedItem = { dedupeKey: string; read: boolean };
+    tdd: `### RED Evidence
 
-export function summarizeDedupedFeed(items: FeedItem[]) {
-  // Last write wins per dedupeKey (stable ordering: later items override earlier ones).
-  const latestReadByKey = new Map<string, boolean>();
-
-  for (const item of items) {
-    latestReadByKey.set(item.dedupeKey, item.read);
-  }
-
-  let unread = 0;
-  for (const read of latestReadByKey.values()) {
-    if (!read) unread += 1;
-  }
-
-  return { uniqueKeys: latestReadByKey.size, unread };
-}
-\`\`\`
-
-### REFACTOR (keep tests green)
+| Slice | Test name | Command | Failure output summary |
+| --- | --- | --- | --- |
+| S-1 (event schema + dedupe) | counts unique keys and unread items | \`\`\`pnpm vitest run tests/unit/dedupe-feed.test.ts\`\`\` | Cannot find module '../notificationFeed' |
+| S-2 (publisher outbox) | publishes event to outbox with dedupe key | \`\`\`pnpm vitest run tests/integration/publisher.test.ts\`\`\` | publishToOutbox is not a function |
+| S-3 (client feed + fallback) | shows notification within 5s via SSE | \`\`\`pnpm playwright test tests/e2e/notification-feed.spec.ts\`\`\` | Element [data-testid="feed-item"] not found |
 
-Keep semantics identical, but make the merge step explicit and easier to unit test in isolation:
+### Acceptance Mapping
 
-\`\`\`typescript
-export type FeedItem = { dedupeKey: string; read: boolean };
+| Slice | Plan task ID | Spec criterion ID |
+| --- | --- | --- |
+| S-1 | T-1 | AC-1, AC-2 |
+| S-2 | T-2 | AC-1 |
+| S-3 | T-3 | AC-1, AC-2, AC-3 |
 
-function mergeLatestByDedupeKey(items: FeedItem[]) {
-  const latestReadByKey = new Map<string, boolean>();
-  for (const item of items) latestReadByKey.set(item.dedupeKey, item.read);
-  return latestReadByKey;
-}
+### Failure Analysis
 
-export function summarizeDedupedFeed(items: FeedItem[]) {
-  const latestReadByKey = mergeLatestByDedupeKey(items);
+| Slice | Expected missing behavior | Actual failure reason |
+| --- | --- | --- |
+| S-1 | notificationFeed module does not exist yet | Module import fails — correct: implementation missing |
+| S-2 | publishToOutbox function not implemented | Function not found — correct: write path missing |
+| S-3 | Feed UI not rendered, SSE not connected | DOM element missing — correct: client component not built |
 
-  let unread = 0;
-  for (const read of latestReadByKey.values()) {
-    if (!read) unread += 1;
-  }
+### GREEN Evidence
 
-  return { uniqueKeys: latestReadByKey.size, unread };
-}
-\`\`\`
+- Full suite command: \`\`\`pnpm vitest run && pnpm playwright test\`\`\`
+- Full suite result: 47 tests passed (3 new + 44 existing), 0 failed, 0 skipped
 
-### Sample terminal output (GREEN)
+### REFACTOR Notes
 
-\`\`\`bash
- RUN  v2.1.0 /Users/dev/app
+- What changed: Extracted \`\`\`mergeLatestByDedupeKey\`\`\` helper from inline loop in \`\`\`summarizeDedupedFeed\`\`\`; moved SSE reconnect logic into \`\`\`useSSEConnection\`\`\` hook.
+- Why: Dedupe merge logic is reused by both publisher and client; reconnect logic was duplicated across components.
+- Behavior preserved: Full suite re-run confirms 47/47 pass after refactor.
 
- ✓ src/notificationFeed.test.ts (1 test) 12ms
+### Traceability
 
- Test Files  1 passed (1)
-      Tests  1 passed (1)
- Tests: 1 passed.
-\`\`\``,
+- Plan task IDs: T-1, T-2, T-3
+- Spec criterion IDs: AC-1, AC-2, AC-3`,
     review: `### Layer 1 — Spec compliance (per-criterion)
 
 | Criterion | Status | Evidence |

package/dist/content/stage-schema.js

@@ -905,6 +905,7 @@ const PLAN = {
         { section: "Task List", required: true, validationRule: "Each task: ID, description, acceptance criterion link, verification command, and effort estimate (S/M/L)." },
         { section: "Acceptance Mapping", required: true, validationRule: "Every spec criterion is covered by at least one task." },
         { section: "Risk Assessment", required: false, validationRule: "If present: per-task or per-wave risk identification with likelihood, impact, and mitigation strategy." },
+        { section: "Boundary Map", required: false, validationRule: "If present: per-wave or per-task interface contracts listing what each task produces (exports) and consumes (imports) from other tasks." },
         { section: "WAIT_FOR_CONFIRM", required: true, validationRule: "Explicit marker present. Status: pending until user approves." }
     ],
     namedAntiPattern: {
@@ -1033,14 +1034,38 @@ const TDD = {
         { name: "Failure-First Thinking", description: "The failing test IS the specification. Until you see the right failure, you do not understand what you are building. Wrong failures are information." },
         { name: "Minimal Viable Change", description: "The best implementation is the smallest one that passes all RED tests. Every extra line is risk. Resist the urge to 'improve while you are here.'" },
         { name: "Regression Paranoia", description: "Assume every change breaks something until the full suite proves otherwise. Partial test runs are lies of omission." },
-        { name: "Refactor-as-Hygiene", description: "Refactoring is not optional cleanup — it is the third leg of TDD. GREEN without REFACTOR accumulates mess. REFACTOR without GREEN breaks things." }
+        { name: "Refactor-as-Hygiene", description: "Refactoring is not optional cleanup — it is the third leg of TDD. GREEN without REFACTOR accumulates mess. REFACTOR without GREEN breaks things." },
+        { name: "Evidence Over Anecdote", description: "Every claim about test state must be backed by captured output. 'It passed' without terminal evidence is not evidence. 'I saw it fail' without the failure output is not RED. Capture commands, outputs, and results — not summaries from memory." }
+    ],
+    reviewSections: [
+        {
+            title: "RED Evidence Audit",
+            evaluationPoints: [
+                "Does every slice have a captured failing test output?",
+                "Does each failure reason match the expected missing behavior (not a typo or config error)?",
+                "Were tests written BEFORE any production code for that slice?",
+                "Does each RED test assert observable behavior, not implementation details?",
+                "Is there a test for each acceptance criterion mapped in the plan?"
+            ],
+            stopGate: true
+        },
+        {
+            title: "GREEN/REFACTOR Audit",
+            evaluationPoints: [
+                "Does GREEN evidence show a FULL suite pass (not partial)?",
+                "Is the GREEN implementation minimal — no features beyond what RED tests require?",
+                "Does the REFACTOR step preserve all existing behavior (no new failures)?",
+                "Are REFACTOR notes documented with rationale?",
+                "Is traceability complete: every change links to plan task ID and spec criterion?"
+            ],
+            stopGate: true
+        }
     ],
-    reviewSections: [],
     completionStatus: ["DONE", "DONE_WITH_CONCERNS", "BLOCKED"],
     crossStageTrace: {
-        readsFrom: [".cclaw/artifacts/05-plan.md", ".cclaw/artifacts/04-spec.md"],
+        readsFrom: [".cclaw/artifacts/05-plan.md", ".cclaw/artifacts/04-spec.md", ".cclaw/artifacts/03-design.md"],
         writesTo: [".cclaw/artifacts/06-tdd.md"],
-        traceabilityRule: "Every RED test traces to a plan task. Every GREEN change traces to a RED test. Every plan task traces to a spec criterion. Evidence chain must be unbroken."
+        traceabilityRule: "Every RED test traces to a plan task. Every GREEN change traces to a RED test. Every plan task traces to a spec criterion. Design decisions inform test strategy. Evidence chain must be unbroken."
     },
     artifactValidation: [
         { section: "RED Evidence", required: true, validationRule: "Failing test output captured per slice." },
@@ -1048,8 +1073,14 @@ const TDD = {
         { section: "Failure Analysis", required: true, validationRule: "Failure reason matches expected missing behavior." },
         { section: "GREEN Evidence", required: true, validationRule: "Full suite pass output captured." },
         { section: "REFACTOR Notes", required: true, validationRule: "What changed, why, behavior preservation confirmed." },
-        { section: "Traceability", required: true, validationRule: "Plan task ID and spec criterion linked." }
+        { section: "Traceability", required: true, validationRule: "Plan task ID and spec criterion linked." },
+        { section: "Verification Ladder", required: false, validationRule: "If present: per-slice verification tier (static, command, behavioral, human) with evidence for highest tier reached." },
+        { section: "Coverage Targets", required: false, validationRule: "If present: per-module or per-code-type coverage thresholds with current values and measurement commands." }
     ],
+    namedAntiPattern: {
+        title: "Code Before Failing Test",
+        description: "Production code written before a failing test is not TDD — it is guessing validated after the fact. Tests written after implementation confirm assumptions, not behavior. If you wrote code first, delete it and start with RED. Delete means delete — not 'keep as reference.' The failing test IS the specification."
+    },
     waveExecutionAllowed: true
 };
 // ---------------------------------------------------------------------------

package/dist/content/templates.js

@@ -278,6 +278,11 @@ Execution rule: complete and verify each wave before starting the next wave.
 |---|---|---|---|---|
 |  |  |  |  |  |
 
+## Boundary Map
+| Task/Wave | Produces (exports) | Consumes (imports from) |
+|---|---|---|
+|  |  |  |
+
 ## WAIT_FOR_CONFIRM
 - Status: pending
 - Confirmed by:
@@ -311,6 +316,17 @@ Execution rule: complete and verify each wave before starting the next wave.
 ## Traceability
 - Plan task IDs:
 - Spec criterion IDs:
+
+
+## Verification Ladder
+| Slice | Tier reached | Evidence |
+|---|---|---|
+| S-1 |  |  |
+
+## Coverage Targets
+| Code type | Target | Current | Command |
+|---|---|---|---|
+|  |  |  |  |
 `,
     "07-review.md": `# Review Artifact
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "cclaw-cli",
-  "version": "0.5.12",
+  "version": "0.5.13",
   "description": "Installer-first flow toolkit for coding agents",
   "type": "module",
   "bin": {