drive: cloud run ce1ea9b8

ops/NEXT.md

@@ -1,87 +1,120 @@
 # NEXT — work package for this tick
 
-**Scope:** Build a minimal agent worker in the SDK. CODE task, SDK-side.
+**Pinned to gate 3** — must not work on any other gate.
 
-This run is pinned to **gate 3** and must not work on any other gate.
+## Scope (quoted from TARGET.md)
 
-## Objective
+Build sub-PR A of the Gate 2 push: a real `hn-monitor` polling runner in the SDK. CODE task, `sdk/src/`-side. This is a scaffolding PR — proof that the workload EXECUTES end-to-end is deliberately deferred to sub-PR B (integration test). Do not conflate the two.
+
+RFC-0001 §3 gate 2 is done when "hn-monitor runs as a relayflow in production, triggered by its real events, with zero bespoke persistence." Every primitive already exists in this repo — event triggers (PR #14, `kernel/relayflowd/tests/event_wake.rs`), the flow spec (`testdata/hn-monitor.flow.yaml`), the poller (`sdk/src/hn-poller.ts`), the agent worker (`sdk/src/worker.ts` from PR #53), a one-shot demo (`sdk/src/demo-hn-monitor.ts`) — but nothing has ever run them together as a continuous workload. This PR fixes that.
 
-Promote the throwaway worker the tests already build into a real SDK component
-that can execute agent steps by running their declared CLI as a subprocess.
+Prior attempt (PR #83, closed): produced a functional runner but was rejected by the swarm on five real findings. Address them in this attempt:
 
-## Context
+1. **Fail-closed on journal errors.** #83's `catch (err) { onPollError(err) }` swallowed EVERY error including `eventSubmit` journal failures — violates covenant 2 (fail-closed) and RFC-0001 §1. Only fetch-level errors (network flakiness, HN API rate limits) may be swallowed; a journal write failure MUST throw and terminate the runner. Split: `try { fetch } catch { onFetchError }` around the network call, `try { eventSubmit } catch { rethrow }` around the journal call.
 
-Nothing in this repo can execute an agent step. Searching for `workerAttach` /
-`step.complete` finds only TESTS (`sdk/tests/live-kernel.test.ts`,
-`journal-client.test.ts`, `journal-client-loopback.ts`) and the protocol
-definitions. `sdk/src/cli/run.ts` only OBSERVES worker leases and waits for one
-that never arrives.
+2. **AgentWorker.close() must release the worker (or explicitly document it does not).** #83 added `await worker.close()` to shutdown but the current `close()` only drains local promises — it does NOT tell the kernel to release the worker registration. Either:
+   - Add a `workerRelease` verb to `sdk/src/protocol.ts` and call it from `close()` (preferred — completes the shutdown contract), OR
+   - Add a one-line comment on `close()` naming exactly what shutdown intentionally does NOT do
 
-The kernel's dispatch, lease and claim machinery is real and tested. The worker
-side of the protocol is simply unimplemented, and that is what blocks gate 2
-("a workload RUNS as a relayflow" — today a run can only be shown CREATED) and
-gate 3 ("every claim/lease/retry served by the kernel").
+3. **Class field declaration order.** #83 declared `private readonly fetcher` AFTER the constructor. Works today because of ES2022 hoisting semantics but breaks silently if someone adds `= someDefault` to a declaration. Declare ALL fields at the top of the class body, before the constructor.
 
-`sdk/tests/live-kernel.test.ts` around the `live-manual-agent` case (line 288)
-shows the whole shape: connect, `hello`, `workerAttach` with pins, receive
-`step.dispatch`, act, complete. The protocol is already proven there.
+4. **Signal handlers must be opt-in via AbortSignal.** #83 registered `SIGTERM`/`SIGINT` handlers on the process directly with no opt-out. A library user embedding this can't cancel one runner without affecting others. Accept `signal?: AbortSignal` in options; the CLI wrapper (sub-PR C) can create + wire a process-signal-driven AbortController.
+
+5. **Test coverage for pollError branch.** #83's tests never asserted the loop survives a fetcher throw AND the loop TERMINATES on a journal throw. Add both cases; without them, someone regresses `onPollError` to a no-op and every test still passes.
+
+## Objective
+
+Add `sdk/src/hn-monitor-runner.ts` composing existing pieces into a continuous runner that addresses all five findings from PR #83.
 
 ## Files in scope
 
-- `sdk/src/worker.ts` — new file, the worker implementation
-- `sdk/src/index.ts` — export the worker
-- `sdk/tests/live-kernel.test.ts` OR a new test file — add a test that runs a
-  real flow with an agent step end to end against a live `relayflowd`, with
-  this worker attached, and asserts the step reaches `done`.
+- `sdk/src/hn-monitor-runner.ts` (NEW) — the runner class
+- `sdk/src/index.ts` — export `HnMonitorRunner`
+- `sdk/src/worker.ts` — MAY modify `close()` per finding #2 above; do NOT rewrite attach/dispatch/complete flow
+- `sdk/src/protocol.ts` — IF adding `workerRelease`, add request/response definitions
+- `sdk/tests/hn-monitor-runner.test.ts` (NEW) — test coverage for ALL cases from finding #5
+- `sdk/package.json` — already has pretest hook that builds kernel
+
+## Definition of Done
+
+ALL of these must hold:
+
+1. `sdk/src/hn-monitor-runner.ts` exists, exports `HnMonitorRunner` from `sdk/src/index.ts`
+
+2. Runner composes the existing pieces:
+   - constructs a `JournalClient` connected to the running `relayflowd` socket
+   - constructs an `AgentWorker` (from `sdk/src/worker.ts`) and calls `workerAttach()` for `agent` steps — attach BEFORE first poll (a run parked because no worker attached is only revived by `run.resume`; the live-kernel suite pins this)
+   - loops: `pollHackerNewsOnce(spec, sink)` → sleep `POLL_INTERVAL_MS` (env-configurable, default 60000 = 60s) → repeat
+   - exit cleanly on `AbortSignal.abort` (drain in-flight steps, close client, release worker per finding #2)
+
+3. Addresses all five findings from PR #83:
+   - Finding #1: Fail-closed on journal errors — fetch errors swallowed, journal errors MUST throw and terminate
+   - Finding #2: `AgentWorker.close()` — either calls `workerRelease` OR has a one-line comment naming what close() intentionally does NOT do
+   - Finding #3: Class field declaration order — ALL fields at top of class body, before constructor
+   - Finding #4: Signal handlers opt-in — accepts `signal?: AbortSignal` in options
+   - Finding #5: Test coverage — see below
+
+4. `sdk/tests/hn-monitor-runner.test.ts` covers ALL of these:
+   - fake fetch + mock journal client → runner submits an event on each tick
+   - abort signal triggers clean shutdown within one tick (worker released or documented)
+   - worker attach happens before first poll
+   - **fetch throw → loop survives** (onPollError called, next tick still runs)
+   - **journal throw → loop TERMINATES** (runner.run() rejects with the error)
+
+5. EVERY new test confirmed to FAIL against current code:
+   - Comment out the source; the test fails
+   - Paste the literal failing output in your summary
 
-## Definition of done
+6. `cd sdk && npm test` green — paste the literal command and output showing test counts
 
-ALL of the following must hold:
+7. PR body explicitly names the non-goals:
+   - Proving the workload actually executes end-to-end (dispatch → step complete) is sub-PR B (integration test with real relayflowd + fake HN fetch + assert step reaches `done`)
+   - CLI wrapper (`flows hn-monitor start`) is sub-PR C
+   - ops/STATE.md gate-2 GREEN declaration is sub-PR D
 
-1. The worker in `sdk/src/worker.ts`, exported from `sdk/src/index.ts`
+8. As your LAST action, run and paste:
+```
+git status --porcelain
+```
 
-2. A test that runs a real flow with an agent step end to end against a live
-   `relayflowd`, with this worker attached, and asserts the step reaches
-   `done`. `sdk/tests/live-kernel.test.ts` already starts a daemon — follow
-   that pattern.
+## Explicit NON-goals for THIS tick (DO NOT TOUCH)
 
-3. **The worker must attach BEFORE the run starts.** A run that finds no worker
-   parks, and attaching afterwards does not re-drive it — `run.resume` is what
-   picks a parked run back up. That contract is pinned in the live-kernel
-   suite; do not fight it.
+- `.github/workflows/*` — no GHA changes
+- `kernel/*` — the kernel side of gate 2 already works via PR #14
+- `workflows/*.yaml` — those are for later sub-PRs
+- `ops/AUTODRIVE_BRIEF.md` — chief owns this file, not the drive loop
+- CLI wrapper — sub-PR C, separate PR
+- end-to-end integration test with real relayflowd — sub-PR B, separate PR
+- ops/STATE.md gate-2 declaration — sub-PR D, separate PR
 
-4. The worker must:
-   - attach for `agent` steps with the pins it holds
-   - on `step.dispatch`, run the step's declared `cli` as a subprocess
-   - report the result back through the existing protocol (`step.complete`, and
-     the failure path when the CLI exits nonzero)
-   - nothing speculative: no retries of its own, no scheduling, no LLM calls.
-     The kernel owns retry and lease policy — do not reimplement it.
+## Current State
 
-5. `cd sdk && npm test` must be green. Run it and paste the literal command and
-   output tail showing test counts.
+**Tests status:**
+- SDK tests: 197 passed (14 files) — all green
+- Kernel tests: 59 passed (8 files) — all green
 
-6. `cd kernel && sh ../ops/cargo.sh test` must be green. Run it and paste the
-   literal command and output tail showing test counts.
+**Existing primitives (ready to compose):**
+- `sdk/src/hn-poller.ts` — `pollHackerNewsOnce()` exists and tested
+- `sdk/src/worker.ts` — `AgentWorker` exists (PR #53)
+- `sdk/src/journal-client.ts` — `JournalClient` exists and tested
+- `sdk/src/demo-hn-monitor.ts` — one-shot demo showing the pattern
+- `testdata/hn-monitor.flow.yaml` — canonical spec exists
+- `kernel/relayflowd/tests/event_wake.rs` — event wake works end-to-end
 
-7. EVERY new test confirmed to FAIL against current code, with the literal
-   failing output quoted in the summary.
+**Missing files:**
+- `sdk/src/hn-monitor-runner.ts` does not exist
+- `sdk/tests/hn-monitor-runner.test.ts` does not exist
 
-8. As your LAST action, run `git status --porcelain` and paste it.
+**Open PRs:** NONE (per ops/STATE.md as of 2026-08-30 02:30)
 
-## Explicitly OUT of scope
+## Blocked or Ready
 
-- LLM steps — not in the gate 3 scope
-- Retry logic in the worker — the kernel owns retry policy
-- Scheduling or lease management — the kernel owns lease policy
-- Optimizations, abstractions, or speculative features
-- Changes to the kernel
-- Changes to existing tests (except adding new test cases)
-- Work on any gate other than gate 3
+READY — all primitives exist, tests are green, no blockers.
 
-## If blocked
+Keep it small and honest:
+- the worker must attach BEFORE the first poll
+- the poller layer handles single-fetch failures with a typed error; the loop just moves to the next tick — but journal errors MUST fail the runner (finding #1)
+- no scheduling logic beyond the sleep (the kernel owns retry and dedupe policy)
+- no LLM calls; the runner is glue, not a reviewer
 
-If gate 3 is genuinely unreachable from the current state, write
-ops/NEEDS_HUMAN.md saying exactly why and still end with ASSESS_DONE. Do not
-silently substitute different work: a run that reports progress on the wrong
-gate is worse than one that reports it is blocked.
+If gate 3 is genuinely unreachable from the current state, write ops/NEEDS_HUMAN.md saying exactly why and still end with ASSESS_DONE. Do not silently substitute different work: a run that reports progress on the wrong gate is worse than one that reports it is blocked.

sdk/package.json

@@ -35,7 +35,7 @@
     "yaml": "^2.5.1"
   },
   "devDependencies": {
-    "@types/node": "^22.7.0",
+    "@types/node": "^22.20.1",
     "typescript": "^5.6.0",
     "vitest": "^2.1.0"
   }

sdk/src/hn-monitor-runner.ts

@@ -0,0 +1,95 @@
+import type { Fetcher } from './hn-poller.js';
+import { pollHackerNewsOnce } from './hn-poller.js';
+import { JournalClient } from './journal-client.js';
+import type { Pins } from './protocol.js';
+import { AgentWorker } from './worker.js';
+
+const DEFAULT_POLL_INTERVAL_MS = 60_000;
+
+export interface HnMonitorRunnerOptions {
+  socketPath: string;
+  spec: unknown;
+  workerId: string;
+  pins: Pins;
+  signal?: AbortSignal;
+  fetcher?: Fetcher;
+  pollIntervalMs?: number;
+  onPollError?: (error: unknown) => void;
+  client?: JournalClient;
+}
+
+/** Continuously submits Hacker News events and services their agent steps. */
+export class HnMonitorRunner {
+  private readonly client: JournalClient;
+  private readonly fetcher: Fetcher | undefined;
+  private readonly onPollError: (error: unknown) => void;
+  private readonly options: HnMonitorRunnerOptions;
+  private readonly pollIntervalMs: number;
+  private readonly worker: AgentWorker;
+
+  constructor(options: HnMonitorRunnerOptions) {
+    this.options = options;
+    this.client = options.client ?? new JournalClient(options.socketPath);
+    this.worker = new AgentWorker(this.client, { workerId: options.workerId, pins: options.pins });
+    this.fetcher = options.fetcher;
+    this.onPollError = options.onPollError ?? (() => undefined);
+    this.pollIntervalMs = options.pollIntervalMs ?? pollIntervalFromEnvironment();
+  }
+
+  async run(): Promise<void> {
+    await this.client.connect();
+    try {
+      await this.worker.attach();
+      while (!this.options.signal?.aborted) {
+        await this.pollOnce();
+        await delay(this.pollIntervalMs, this.options.signal);
+      }
+    } finally {
+      await this.worker.close();
+      this.client.close();
+    }
+  }
+
+  private async pollOnce(): Promise<void> {
+    let journalFailed = false;
+    try {
+      await pollHackerNewsOnce(this.options.spec, {
+        eventSubmit: async (spec, event) => {
+          try {
+            return await this.client.eventSubmit(spec, event);
+          } catch (error) {
+            journalFailed = true;
+            throw error;
+          }
+        },
+      }, { fetcher: this.fetcher });
+    } catch (error) {
+      if (journalFailed) throw error;
+      this.onPollError(error);
+    }
+  }
+}
+
+function pollIntervalFromEnvironment(): number {
+  const configured = process.env.POLL_INTERVAL_MS;
+  if (configured === undefined) return DEFAULT_POLL_INTERVAL_MS;
+  const interval = Number(configured);
+  if (!Number.isFinite(interval) || interval < 0) {
+    throw new Error(`POLL_INTERVAL_MS must be a non-negative number, received ${configured}`);
+  }
+  return interval;
+}
+
+function delay(ms: number, signal?: AbortSignal): Promise<void> {
+  if (signal?.aborted) return Promise.resolve();
+  return new Promise((resolve) => {
+    const timer = setTimeout(finish, ms);
+    signal?.addEventListener('abort', finish, { once: true });
+
+    function finish(): void {
+      clearTimeout(timer);
+      signal?.removeEventListener('abort', finish);
+      resolve();
+    }
+  });
+}

sdk/src/index.ts

@@ -113,6 +113,7 @@ export { JOURNAL_WRITE_FAILED, PROTOCOL_VERSION } from './protocol.js';
 
 export { JournalClient, type JournalClientOptions } from './journal-client.js';
 export { AgentWorker, type AgentWorkerOptions } from './worker.js';
+export { HnMonitorRunner, type HnMonitorRunnerOptions } from './hn-monitor-runner.js';
 
 export {
   validateWorkPackage,

sdk/src/worker.ts

@@ -18,6 +18,7 @@ interface CliResult {
 /** Executes dispatched agent steps using their declared CLI. */
 export class AgentWorker extends EventEmitter {
   private attached = false;
+  private readonly active = new Set<Promise<void>>();
 
   constructor(
     private readonly client: JournalClient,
@@ -38,14 +39,20 @@ export class AgentWorker extends EventEmitter {
     }
   }
 
-  close(): void {
+  async close(): Promise<void> {
     this.client.off('step.dispatch', this.onDispatch);
     this.attached = false;
+    await Promise.allSettled(this.active);
+    // Intentionally does not release the kernel worker registration: protocol v0 has no worker.release verb.
   }
 
   private readonly onDispatch = (dispatch: StepDispatchEvent): void => {
     if (dispatch.step_type !== 'agent') return;
-    void this.execute(dispatch).catch((error: unknown) => this.emit('error', error));
+    const execution = this.execute(dispatch);
+    this.active.add(execution);
+    void execution
+      .catch((error: unknown) => this.emit('error', error))
+      .finally(() => this.active.delete(execution));
   };
 
   private async execute(dispatch: StepDispatchEvent): Promise<void> {