fix: close Wow Wave sub-issue gaps (#1444 + build drift)

apps/mcp-server/src/rules/rule-insights.service.spec.ts

@@ -215,7 +215,7 @@ describe('RuleInsightsService', () => {
             generatedRule: true,
             baselineFailureRate: 0.25,
             currentFailureRate: 0.05,
-          } as unknown as RuleStats,
+          },
         };
 
         const result = service.generateInsights(stats, [], NOW);
@@ -235,7 +235,7 @@ describe('RuleInsightsService', () => {
             generatedRule: true,
             baselineFailureRate: 0.2,
             currentFailureRate: 0.22,
-          } as unknown as RuleStats,
+          },
         };
 
         const result = service.generateInsights(stats, [], NOW);
@@ -252,7 +252,7 @@ describe('RuleInsightsService', () => {
             generatedRule: true,
             baselineFailureRate: 0.3,
             currentFailureRate: 0.2,
-          } as unknown as RuleStats,
+          },
         };
 
         const result = service.generateInsights(stats, [], NOW);

apps/mcp-server/src/rules/rule-insights.service.ts

@@ -139,21 +139,15 @@ export class RuleInsightsService {
     const scores: EffectivenessScore[] = [];
 
     for (const [name, stat] of entries) {
-      const extended = stat as RuleStats & {
-        generatedRule?: boolean;
-        baselineFailureRate?: number;
-        currentFailureRate?: number;
-      };
-
-      if (!extended.generatedRule) continue;
+      if (!stat.generatedRule) continue;
       if (
-        typeof extended.baselineFailureRate !== 'number' ||
-        typeof extended.currentFailureRate !== 'number'
+        typeof stat.baselineFailureRate !== 'number' ||
+        typeof stat.currentFailureRate !== 'number'
       )
         continue;
 
-      const baseline = extended.baselineFailureRate;
-      const current = extended.currentFailureRate;
+      const baseline = stat.baselineFailureRate;
+      const current = stat.currentFailureRate;
       const reductionPercent = baseline > 0 ? ((baseline - current) / baseline) * 100 : 0;
 
       let verdict: EffectivenessScore['verdict'];

apps/mcp-server/src/rules/rule-tracker.spec.ts

@@ -189,4 +189,94 @@ describe('RuleTracker', () => {
       expect(loaded.getStats()).toEqual({});
     });
   });
+
+  describe('markRuleAsGenerated (#1444)', () => {
+    it('should tag an existing rule as auto-generated', () => {
+      tracker.trackRuleUsage(['auto-bash-guard']);
+      tracker.markRuleAsGenerated('auto-bash-guard');
+
+      const stats = tracker.getStats();
+      expect(stats['auto-bash-guard'].generatedRule).toBe(true);
+    });
+
+    it('should create a fresh stat entry when marking an unseen rule', () => {
+      tracker.markRuleAsGenerated('fresh-rule');
+
+      const stats = tracker.getStats();
+      expect(stats['fresh-rule']).toBeDefined();
+      expect(stats['fresh-rule'].generatedRule).toBe(true);
+      expect(stats['fresh-rule'].count).toBe(0);
+    });
+
+    it('should ignore empty rule name', () => {
+      tracker.markRuleAsGenerated('');
+
+      expect(tracker.getStats()).toEqual({});
+    });
+
+    it('should not wipe generatedRule flag on subsequent trackRuleUsage', () => {
+      tracker.markRuleAsGenerated('auto-rule');
+      tracker.trackRuleUsage(['auto-rule']);
+      tracker.trackRuleUsage(['auto-rule']);
+
+      const stats = tracker.getStats();
+      expect(stats['auto-rule'].generatedRule).toBe(true);
+      expect(stats['auto-rule'].count).toBe(2);
+    });
+  });
+
+  describe('recordFailureRate (#1444)', () => {
+    it('should record baseline failure rate for a generated rule', () => {
+      tracker.markRuleAsGenerated('auto-rule');
+      tracker.recordFailureRate('auto-rule', 0.25, 'baseline');
+
+      const stats = tracker.getStats();
+      expect(stats['auto-rule'].baselineFailureRate).toBe(0.25);
+    });
+
+    it('should record current failure rate for a generated rule', () => {
+      tracker.markRuleAsGenerated('auto-rule');
+      tracker.recordFailureRate('auto-rule', 0.05, 'current');
+
+      const stats = tracker.getStats();
+      expect(stats['auto-rule'].currentFailureRate).toBe(0.05);
+    });
+
+    it('should overwrite current failure rate on repeat calls', () => {
+      tracker.markRuleAsGenerated('auto-rule');
+      tracker.recordFailureRate('auto-rule', 0.2, 'current');
+      tracker.recordFailureRate('auto-rule', 0.1, 'current');
+
+      const stats = tracker.getStats();
+      expect(stats['auto-rule'].currentFailureRate).toBe(0.1);
+    });
+
+    it('should be a no-op for rules not marked as generated', () => {
+      tracker.trackRuleUsage(['manual-rule']);
+      tracker.recordFailureRate('manual-rule', 0.5, 'baseline');
+
+      const stats = tracker.getStats();
+      expect(stats['manual-rule'].baselineFailureRate).toBeUndefined();
+    });
+
+    it('should be a no-op for unknown rules', () => {
+      tracker.recordFailureRate('unknown-rule', 0.5, 'baseline');
+
+      expect(tracker.getStats()).toEqual({});
+    });
+
+    it('should persist failure rates through save/load cycle', async () => {
+      vi.mocked(existsSync).mockReturnValue(true);
+      tracker.markRuleAsGenerated('auto-rule');
+      tracker.recordFailureRate('auto-rule', 0.3, 'baseline');
+      tracker.recordFailureRate('auto-rule', 0.05, 'current');
+
+      await tracker.save();
+
+      const written = JSON.parse(vi.mocked(fs.writeFile).mock.calls[0][1] as string);
+      expect(written['auto-rule'].generatedRule).toBe(true);
+      expect(written['auto-rule'].baselineFailureRate).toBe(0.3);
+      expect(written['auto-rule'].currentFailureRate).toBe(0.05);
+    });
+  });
 });

apps/mcp-server/src/rules/rule-tracker.ts

@@ -5,6 +5,22 @@ import { dirname } from 'path';
 export interface RuleStats {
   count: number;
   lastUsed: number;
+  /**
+   * True when the rule was generated by the self-evolving pipeline
+   * (#1444). Used to separate auto-generated rules from manual rules in
+   * effectiveness reports.
+   */
+  generatedRule?: boolean;
+  /**
+   * Baseline failure rate captured when a generated rule was first
+   * introduced. Combined with {@link currentFailureRate} to compute
+   * effectiveness reduction percent.
+   */
+  baselineFailureRate?: number;
+  /**
+   * Latest observed failure rate for a generated rule.
+   */
+  currentFailureRate?: number;
 }
 
 export interface RuleEffectivenessReport {
@@ -46,6 +62,37 @@ export class RuleTracker {
     }
   }
 
+  /**
+   * Mark a rule as auto-generated by the self-evolving pipeline (#1444).
+   * Creates the stat entry if the rule has never been tracked.
+   */
+  markRuleAsGenerated(ruleName: string): void {
+    if (!ruleName) return;
+    if (!this.stats[ruleName]) {
+      this.stats[ruleName] = { count: 0, lastUsed: Date.now() };
+    }
+    this.stats[ruleName].generatedRule = true;
+  }
+
+  /**
+   * Record a baseline or current failure rate for a generated rule
+   * (#1444). Baseline is captured once when the rule is introduced;
+   * current is updated as new execution data arrives.
+   *
+   * A rule must be marked as generated (see {@link markRuleAsGenerated})
+   * before failure rates are recorded — otherwise the call is a no-op.
+   */
+  recordFailureRate(ruleName: string, rate: number, kind: 'baseline' | 'current'): void {
+    if (!ruleName) return;
+    const stat = this.stats[ruleName];
+    if (!stat || !stat.generatedRule) return;
+    if (kind === 'baseline') {
+      stat.baselineFailureRate = rate;
+    } else {
+      stat.currentFailureRate = rate;
+    }
+  }
+
   getStats(): Record<string, RuleStats> {
     return JSON.parse(JSON.stringify(this.stats));
   }

packages/claude-code-plugin/README.md

@@ -2,7 +2,7 @@
 
 # CodingBuddy Claude Code Plugin
 
-> Version 5.4.0
+> Version 5.4.1
 
 Multi-AI Rules for consistent coding practices - PLAN/ACT/EVAL workflow, specialist agents, and reusable skills for systematic development.
 

packages/claude-code-plugin/namespace-manifest.json

@@ -34,5 +34,5 @@
       "namespaced": "codingbuddy:plan"
     }
   ],
-  "generatedAt": "2026-04-06T09:54:05.647Z"
+  "generatedAt": "2026-04-10T09:03:53.981Z"
 }

packages/claude-code-plugin/scripts/build.ts

@@ -119,14 +119,18 @@ The plugin works standalone with core features. MCP integration unlocks the full
 | Feature                              | Standalone | With MCP |
 | ------------------------------------ | :--------: | :------: |
 | PLAN/ACT/EVAL/AUTO keyword triggers  |     ✅     |    ✅    |
+| Staged planning (Discover→Design→Plan) |   ✅     |    ✅    |
+| Clarification gate (ambiguous prompts) |   ✅     |    ✅    |
+| Permission forecast (prompt-aware)   |     ✅     |    ✅    |
+| Council scene (agent eye glyphs)     |   ✅ ¹    |    ✅    |
 | Specialist agent prompts             |   ⚠️ ¹    |    ✅    |
 | Slash commands (codingbuddy:\\*)      |     ✅     |    ✅    |
 | Dynamic checklists                   |     —      |    ✅    |
 | Context persistence across modes     |     —      |    ✅    |
 | Rule search & impact reports         |     —      |    ✅    |
 | Session briefings & recovery         |     —      |    ✅    |
 
-> ¹ Standalone specialist prompts require a local \`.ai-rules/\` directory. Run \`npx codingbuddy-rules init\` to scaffold it. See [#1216](https://github.com/JeremyDev87/codingbuddy/issues/1216) for the standalone enhancement roadmap.
+> ¹ Standalone features that read agent definitions require a local \`.ai-rules/\` directory. Run \`npx codingbuddy-rules init\` to scaffold it. See [#1216](https://github.com/JeremyDev87/codingbuddy/issues/1216) for the standalone enhancement roadmap.
 
 ## MCP Integration (Recommended)