feat: add tool allow/deny profiles with per-agent and per-provider filtering

Implements configurable tool filtering with four built-in profiles
(minimal, messaging, coding, full), global and per-agent/per-provider
allow/deny lists with glob pattern support, and defense-in-depth
enforcement at both tool listing and execution time.

New: src/tools/policy.ts (ToolPolicy engine), src/tools/policy.test.ts (37 tests)
Modified: config schema, tool registry, tool executor, NativeAgent,
AgentOrchestrator, daemon wiring, gateway tool handler, test mocks

src/backends/native/orchestrator.ts

@@ -4,6 +4,7 @@ import type { Session } from '../../session/index.js';
 import type { ToolRegistry } from '../../tools/registry.js';
 import type { ToolExecutor } from '../../tools/executor.js';
 import type { MemoryStore } from '../../memory/store.js';
+import type { ToolPolicyContext } from '../../tools/policy.js';
 import { NativeAgent } from './agent.js';
 import type { ToolUseEvent } from './agent.js';
 import { shouldCompact } from '../../context/tokens.js';
@@ -87,6 +88,8 @@ export interface OrchestratorConfig {
   contextWindow?: number;
   /** Optional memory store for injecting persistent memory into the system prompt. */
   memoryStore?: MemoryStore;
+  /** Policy context for tool filtering (agent tier, provider). */
+  toolPolicyContext?: ToolPolicyContext;
 }
 
 // ── AgentOrchestrator ─────────────────────────────────────────────────
@@ -134,6 +137,7 @@ export class AgentOrchestrator {
       toolExecutor: config.toolExecutor,
       maxIterations: config.maxIterations,
       onToolUse: config.onToolUse,
+      toolPolicyContext: config.toolPolicyContext,
     });
 
     // Set the primary tier on the agent
@@ -174,9 +178,10 @@ export class AgentOrchestrator {
       maxTokens: request.maxTokens,
     };
 
-    // Optionally include tools from the registry
+    // Optionally include tools from the registry (filtered by policy)
     if (request.tools && this._toolRegistry) {
-      chatRequest.tools = this._toolRegistry.toAnthropicFormat();
+      const policyContext = this._agent.getToolPolicyContext();
+      chatRequest.tools = this._toolRegistry.filteredToAnthropicFormat(policyContext);
     }
 
     const response = await this._modelRouter.chat(chatRequest, tier);