feat(ai): policy is read from the owning agent (Phase 4)

Until now AiPolicy lived as a user-global setting consulted for every
AI action. With agents as the principal unit of AI behavior, policy
belongs on the agent — different agents can be aggressive about tasks
but conservative about calendar edits, etc.

Webapp (tools/executor.ts):
- When an AI actor invokes a tool, the executor looks up the owning
  agent via getAgent(actor.principalId) and passes agent.policy into
  resolvePolicy. Falls back to DEFAULT_AI_POLICY when the agent record
  is missing (legacy write, deleted agent, race) so no tool call can
  silently bypass the propose/deny path.
- resolvePolicy already accepted an optional policy arg, so the call
  site change is a single line plus the agent load.

Server (mana-ai):
- ServerAgent gains an optional policy field, projected off the same
  plaintext JSONB that the webapp writes.
- Tick loop filters AI_AVAILABLE_TOOLS through filterToolsByAgentPolicy
  before passing them to the planner prompt. Resolution order mirrors
  the webapp: tools[name] → defaultsByModule → defaultForAi; 'deny'
  drops the tool so the LLM never even sees it.

Phase 5 will surface a per-agent policy editor on the agent-detail
UI. Until then all agents inherit DEFAULT_AI_POLICY (baked in during
createAgent), which means no behavior change for existing users —
every tool that was 'propose' before is still 'propose' now, just
reached via agent.policy instead of the user-level singleton.

Tests: mana-ai 41/41, webapp svelte-check clean.

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

services/mana-ai/src/cron/tick.ts

@@ -240,7 +240,7 @@ async function planOneMission(
 	const input: AiPlanInput = {
 		mission,
 		resolvedInputs,
-		availableTools: AI_AVAILABLE_TOOLS,
+		availableTools: filterToolsByAgentPolicy(AI_AVAILABLE_TOOLS, agent),
 	};
 	const messages = withAgentContext(buildPlannerPrompt(input), agent);
 	const result = await planner.complete(messages);
@@ -265,6 +265,31 @@ async function planOneMission(
  * Ciphertext fields (`enc:1:…`) are intentionally skipped — the server
  * doesn't hold the decrypt key; the foreground runner handles those.
  */
+/**
+ * Drop tools that the agent's policy denies, so the Planner never even
+ * sees a tool it can't use. Tools with policy `propose` stay in the
+ * allowlist (they just get proposed rather than auto-run on the user's
+ * device), and `auto` tools stay too. A missing policy or missing
+ * agent leaves the list unchanged.
+ *
+ * Resolution order matches the webapp's `resolvePolicy`:
+ *   tools[name] ?? defaultsByModule[tool.module] ?? defaultForAi
+ */
+function filterToolsByAgentPolicy(
+	tools: readonly import('@mana/shared-ai').AvailableTool[],
+	agent: ServerAgent | null
+): import('@mana/shared-ai').AvailableTool[] {
+	if (!agent?.policy) return tools as import('@mana/shared-ai').AvailableTool[];
+	const policy = agent.policy;
+	return tools.filter((t) => {
+		const byTool = policy.tools[t.name];
+		if (byTool) return byTool !== 'deny';
+		const byModule = policy.defaultsByModule?.[t.module];
+		if (byModule) return byModule !== 'deny';
+		return policy.defaultForAi !== 'deny';
+	});
+}
+
 function withAgentContext(messages: PlannerMessages, agent: ServerAgent | null): PlannerMessages {
 	if (!agent) return messages;
 

services/mana-ai/src/db/agents-projection.ts

@@ -19,6 +19,7 @@
 
 import type { Sql } from './connection';
 import { withUser } from './connection';
+import type { AiPolicy } from '@mana/shared-ai';
 
 export interface ServerAgent {
 	id: string;
@@ -31,6 +32,10 @@ export interface ServerAgent {
 	 *  only injects plaintext. */
 	systemPrompt?: string;
 	memory?: string;
+	/** Per-tool auto/propose/deny — drives the server-side tool
+	 *  allowlist when Phase 4 wiring is complete. Plaintext. Undefined
+	 *  on legacy agent records (pre-Phase-2 writes). */
+	policy?: AiPolicy;
 	state: 'active' | 'paused' | 'archived';
 	maxConcurrentMissions: number;
 	maxTokensPerDay?: number;
@@ -238,6 +243,10 @@ function toServerAgent(row: SnapshotRow): ServerAgent {
 		role: String(r.role ?? ''),
 		systemPrompt: typeof r.systemPrompt === 'string' ? r.systemPrompt : undefined,
 		memory: typeof r.memory === 'string' ? r.memory : undefined,
+		policy:
+			r.policy && typeof r.policy === 'object' && !Array.isArray(r.policy)
+				? (r.policy as AiPolicy)
+				: undefined,
 		state: (r.state as ServerAgent['state']) ?? 'active',
 		maxConcurrentMissions: Number(r.maxConcurrentMissions ?? 1),
 		maxTokensPerDay: typeof r.maxTokensPerDay === 'number' ? r.maxTokensPerDay : undefined,