✅ test(claw-manager): add 7 missing unit tests + document analytics

New tests (29 → 36):
- denyEscalation: resume with inbox message, non-existent claw, non-escalated
- stop conditions: on_error (first failure), idle:N (consecutive idle cycles)
- config hot-reload: updateClawConfig in-memory, no-op for unknown

CLAUDE.md: add AnalyticsPage entry, bump test count to 26,750+

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: ersinkoc

CLAUDE.md

@@ -22,7 +22,8 @@ packages/
 - **User Extensions**: Native tool bundles (JS code, triggers, services) in `packages/gateway/src/services/extension-service.ts`. DB table: `user_extensions`. API: `/extensions`
 - **Skills (AgentSkills.io)**: Open standard SKILL.md format for agent instructions. Parser: `packages/gateway/src/services/agentskills-parser.ts`. Format field: `'ownpilot' | 'agentskills'`
 - **Edge/IoT**: MQTT broker (Mosquitto) integration for edge device management. Types: `packages/core/src/edge/`. Service: `packages/gateway/src/services/edge-service.ts`. Routes: `/api/v1/edge`
-- **Test framework**: Vitest across all packages. 26,700+ tests total (gateway: 16,500+; core: 9,714; cli: 340; ui: 141). 550 test files
+- **Test framework**: Vitest across all packages. 26,750+ tests total (gateway: 16,550+; core: 9,714; cli: 340; ui: 141). 550 test files
+- **Analytics Page**: `packages/ui/src/pages/AnalyticsPage.tsx` — recharts-powered dashboard at `/analytics`. 6 KPI cards, cost/token area+bar charts, provider donut, agent distribution bar, claw mode/state donuts, task/habit radial gauges, daily requests line chart, claw runtime summary grid, personal data overview. Period toggle (7d/30d). Uses `costsApi.usage()`, `costsApi.getBreakdown()`, `clawsApi.stats()`, `summaryApi.get()` + agent list endpoints
 - **Autonomous Agent Runners**: Shared utilities in `packages/gateway/src/services/agent-runner-utils.ts` — `createConfiguredAgent()`, `registerAllToolSources()`, `resolveProviderAndModel()`, `executeAgentPipeline()`, `calculateExecutionCost()`, `createToolCallCollector()`, `resolveToolFilter()`, `createCancellationPromise()`
 - **Habit Tracking**: 8 AI tools in `packages/gateway/src/tools/habit-tools.ts`, DB repo in `db/repositories/habits.ts` (645 lines), REST API in `routes/productivity.ts`, HabitsPage UI with streak heatmap
 - **Utilities**: `TTLCache<K,V>` in `packages/gateway/src/utils/ttl-cache.ts` — generic cache with auto-prune. `chat-post-processor.ts` in `assistant/` — extracted from conversation-service

packages/gateway/src/services/claw-manager.test.ts

@@ -369,6 +369,41 @@ describe('ClawManager', () => {
       const approved = await manager.approveEscalation('claw-1');
       expect(approved).toBe(false);
     });
+
+    it('should deny escalation and resume with inbox message', async () => {
+      const repo = setupRepo(makeConfig());
+      await manager.startClaw('claw-1', 'user-1');
+
+      await manager.requestEscalation('claw-1', {
+        id: 'esc-2',
+        type: 'tool_access',
+        reason: 'Need shell',
+        requestedAt: new Date(),
+      });
+      expect(manager.getSession('claw-1')?.state).toBe('escalation_pending');
+
+      const denied = await manager.denyEscalation('claw-1', 'Too risky');
+      expect(denied).toBe(true);
+      expect(manager.getSession('claw-1')?.state).toBe('running');
+      expect(manager.getSession('claw-1')?.pendingEscalation).toBeNull();
+      expect(manager.getSession('claw-1')?.inbox).toContainEqual(
+        expect.stringContaining('ESCALATION_DENIED')
+      );
+      expect(repo.appendToInbox).toHaveBeenCalled();
+    });
+
+    it('should return false when denying non-existent claw', async () => {
+      const denied = await manager.denyEscalation('claw-99');
+      expect(denied).toBe(false);
+    });
+
+    it('should return false when denying non-escalated claw', async () => {
+      setupRepo(makeConfig());
+      await manager.startClaw('claw-1', 'user-1');
+
+      const denied = await manager.denyEscalation('claw-1');
+      expect(denied).toBe(false);
+    });
   });
 
   describe('stop conditions', () => {
@@ -398,6 +433,53 @@ describe('ClawManager', () => {
 
       expect(manager.getSession('claw-1')).toBeNull();
     });
+
+    it('should stop on on_error condition when cycle fails', async () => {
+      const config = makeConfig({ mode: 'continuous', stopCondition: 'on_error' });
+      setupRepo(config);
+      mockRunCycle.mockResolvedValue(makeCycleResult({ success: false, error: 'Something broke' }));
+
+      await manager.startClaw('claw-1', 'user-1');
+      await vi.advanceTimersByTimeAsync(600);
+
+      // on_error stops after first failure
+      expect(manager.getSession('claw-1')).toBeNull();
+    });
+
+    it('should stop on idle:N condition after N idle cycles', async () => {
+      const config = makeConfig({ mode: 'continuous', stopCondition: 'idle:2' });
+      setupRepo(config);
+      // Cycle returns 0 tool calls (idle)
+      mockRunCycle.mockResolvedValue(makeCycleResult({ toolCalls: [] }));
+
+      await manager.startClaw('claw-1', 'user-1');
+      // First idle cycle
+      await vi.advanceTimersByTimeAsync(600);
+      expect(manager.getSession('claw-1')).not.toBeNull();
+      // Second idle cycle → stop
+      await vi.advanceTimersByTimeAsync(5100);
+      expect(manager.getSession('claw-1')).toBeNull();
+    });
+  });
+
+  describe('config hot-reload', () => {
+    it('should update in-memory config via updateClawConfig', async () => {
+      const config = makeConfig({ mode: 'continuous' });
+      setupRepo(config);
+      await manager.startClaw('claw-1', 'user-1');
+
+      const updated = { ...config, mode: 'interval' as const, intervalMs: 60_000 };
+      manager.updateClawConfig('claw-1', updated);
+
+      expect(manager.getSession('claw-1')?.config.mode).toBe('interval');
+      expect(manager.getSession('claw-1')?.config.intervalMs).toBe(60_000);
+    });
+
+    it('should no-op for unknown claw', () => {
+      const config = makeConfig();
+      // Should not throw
+      manager.updateClawConfig('claw-99', config);
+    });
   });
 
   describe('resource limits', () => {