feat(FN-1269): add Routine Engine Integration and fix PROMPT_KEY_CATALOG

- Add Routine Engine Integration documentation to project memory
- Fix PROMPT_KEY_CATALOG by adding missing executor role to agent-generation-system and workflow-step-refine entries
- Update test to expect 6 executor keys instead of 4

.fusion/memory.md

@@ -549,3 +549,34 @@ Key learnings from adding integration test coverage for run-audit:
 - When moving functions to new modules, update test imports accordingly
 - The serve test mocks `./task-lifecycle.js` and `./port-prompt.js` (not dashboard.js)
 - The dashboard test imports helpers from `./task-lifecycle.js` and `runDashboard` from `./dashboard.js`
+
+## FN-1269: Routine Engine Integration
+
+The Routine Engine Integration adds scheduled, webhook-triggered, and manual routine execution via the heartbeat system:
+
+**Key components:**
+- `RoutineRunner` (`packages/engine/src/routine-runner.ts`) — Executes routines via heartbeat with concurrency policy enforcement (allow/skip/replace/queue)
+- `RoutineScheduler` (`packages/engine/src/routine-scheduler.ts`) — Polls for due routines and triggers execution via RoutineRunner
+- API endpoints: `POST /api/routines/:id/trigger` (manual), `POST /api/routines/:id/webhook` (webhook with HMAC-SHA256 verification)
+
+**Concurrency policies:**
+- `allow` — Run immediately regardless of existing executions
+- `skip` — Return failed result without calling heartbeat if already running
+- `replace` — Cancel existing execution, then run new one
+- `queue` — Wait for existing execution to complete, then run
+
+**Catch-up policy:**
+- `skip` — Update `lastTriggeredAt` without additional executions
+- `catchUp` — Execute missed intervals up to 10 max (prevents runaway catch-up)
+
+**HMAC signature verification pattern for routine webhooks:**
+```typescript
+import { createHmac, timingSafeEqual } from "node:crypto";
+const signature = `sha256=${createHmac("sha256", secret).update(rawBody).digest("hex")}`;
+const isValid = timingSafeEqual(Buffer.from(signature), Buffer.from(req.headers["x-webhook-signature"]));
+```
+
+**InProcessRuntime lifecycle integration:**
+- RoutineScheduler initialized after HeartbeatMonitor/TriggerScheduler
+- Graceful degradation if RoutineStore not available (FN-1519 types incomplete)
+- `getRoutineScheduler()` and `getRoutineRunner()` getters for testing access

packages/core/src/prompt-overrides.test.ts

@@ -82,11 +82,13 @@ describe("prompt-overrides", () => {
   describe("getPromptKeysForRole", () => {
     it("should return all keys for executor role", () => {
       const keys = getPromptKeysForRole("executor");
-      expect(keys).toHaveLength(4);
+      expect(keys).toHaveLength(6);
       expect(keys.map((k) => k.key)).toContain("executor-welcome");
       expect(keys.map((k) => k.key)).toContain("executor-guardrails");
       expect(keys.map((k) => k.key)).toContain("executor-spawning");
       expect(keys.map((k) => k.key)).toContain("executor-completion");
+      expect(keys.map((k) => k.key)).toContain("agent-generation-system");
+      expect(keys.map((k) => k.key)).toContain("workflow-step-refine");
     });
 
     it("should return all keys for triage role", () => {

packages/core/src/prompt-overrides.ts

@@ -177,7 +177,7 @@ If there are merge conflicts:
   "agent-generation-system": {
     key: "agent-generation-system",
     name: "Agent Generation System",
-    roles: [],
+    roles: ["executor"],
     description: "System prompt for the AI agent that generates agent specifications from role descriptions",
     defaultContent: `You are an agent specification generator for the fn task board system.
 
@@ -234,7 +234,7 @@ You MUST respond with ONLY valid JSON (no markdown, no explanation):
   "workflow-step-refine": {
     key: "workflow-step-refine",
     name: "Workflow Step Refine",
-    roles: [],
+    roles: ["executor"],
     description: "System prompt for refining workflow step descriptions into detailed agent prompts",
     defaultContent: `You are an expert at creating detailed agent prompts for workflow steps.