feat(policy): enforce truthfulness and autonomy guardrails

Add runtime truthfulness modes and autonomy-level tool gating with audit metadata for overrides/denials.

Wire policy through prompt assembly, tool execution context, and daemon/gateway agent paths; update tests and planning state for Phase 3 PR #2 completion.

src/tools/integration.test.ts

@@ -44,6 +44,7 @@ describe('Tool integration (end-to-end)', () => {
       systemPrompt: 'You have tools.',
       toolRegistry: registry,
       toolExecutor: executor,
+      toolPolicyContext: { autonomyLevel: 'autonomous' },
     });
 
     const result = await agent.process('run echo integration_test');
@@ -89,9 +90,10 @@ describe('Tool integration (end-to-end)', () => {
 
     const agent = new NativeAgent({
       modelClient: mockClient,
-      systemPrompt: 'You have file tools.',
+      systemPrompt: 'You have tools.',
       toolRegistry: registry,
       toolExecutor: executor,
+      toolPolicyContext: { autonomyLevel: 'autonomous' },
     });
 
     try {
@@ -141,6 +143,7 @@ describe('Tool integration (end-to-end)', () => {
       systemPrompt: 'You have tools.',
       toolRegistry: registry,
       toolExecutor: executor,
+      toolPolicyContext: { autonomyLevel: 'autonomous' },
     });
 
     const result = await agent.process('verify tool results');