Files
PaperClipAI/server/src/__tests__/durable-chat-wakeup.test.ts
T
DottaandPaperclip f0877a9c7c fix: recover native work and requalify saved computers (#17724)
<!-- entire-trail-link-start -->
https://entire.com/gh/paperclipai/-/trails/32/changes/paperclip/32
<!-- entire-trail-link-end -->

## Thinking Path

> - Paperclip manages AI agents and preserves their work after an
interruption.
> - Native Stop must confirm termination and save the workspace.
> - Task and run views must show the current recovery receipt.
> - A new external reply can arrive while that cleanup is pending.
> - The reply must return to its original chat owner after cleanup.
> - A successful fresh retry also becomes the task’s bound workspace for
its next ordinary turn.
> - Managed computer updates can invalidate a saved snapshot
qualification.
> - This pull request fixes those instructions, acknowledgements,
handoffs and the existing native-agent Test path.
> - The operator can continue with new input while the stopped turn
remains preserved.

## Linked Issues or Issue Description

**What happened?**

Stopped native runs could keep the initial workspace-repair instruction
after files were recovered. Native Stop could also leave the shared
cancellation state at requested after remote termination was confirmed.
A new GitHub reply saved during cleanup could then remain queued because
cleanup did not ask its original inbound owner to reconsider the
receipt. After a managed snapshot update, an existing native agent could
also need qualification while the generic environment Test required
instance-admin access.

**Expected behavior**

Show the current saved-work state. Acknowledge Stop only after exact
termination proof. Let an authorized operator use the existing
saved-agent Run test to qualify its current computer. After cleanup,
recheck the saved reply's current authority and admit it once as new
input. Preserve the stopped run and uncertain external effects.

**Steps to reproduce**

1. Stop an active native turn before it returns a semantic result.
2. Send a new authorized external reply while cleanup is pending.
3. Let workspace recovery and provider cleanup finish. Inspect the run
instruction, Stop acknowledgement and queued reply.

Related work: #17594 preserves native processes through controller
outages. #17401 addresses a separate bootstrap failure. This pull
request does not replace those fixes.

## What Changed

- Use the exact run's workspace recovery receipt for task and run
guidance, including the Activity ledger.
- Retire an obsolete exact-session notice through the existing bounded
reconciler. Preserve the separate no-replay hold.
- Write the shared native Stop acknowledgement only after the audited
Stop intent and every exact remote lease receipt match.
- Delegate saved external input to the live inbound owner. Recheck its
source, actor, endpoint, principal and conversation generation. Promote
the same receipt in the transaction that creates its new run.
- Use that owner from the cleanup callback and existing stranded-queue
catch-up after a lost callback or restart. A denied source cannot abort
other items.
- Reuse the current private task directory after a successful fresh
retry. Re-read the company/task/workspace binding and validate the
canonical stable or attempt namespace without recreating missing files;
keep a later fresh retry separate.
- Extend the existing agent-scoped qualification owner to native agents.
The clean saved-agent Test qualifies its exact computer before runtime
testing. Reuse a current proof; require a real probe and verified
cleanup for a stale proof. Preserve draft tests and metadata-only
external/pool previews.
- Update execution semantics, capability maps and the recovery
verification recipe. Add a production-page Storybook interaction for the
normal native Test.

## Verification

- **Red:** persisted recovery states returned stale instructions; an
exact terminated native lease left shared Stop requested; the real
GitHub subscribed-message fixture left zero successors after cleanup.
The controlled baseline restored the previous production source,
reproduced the failure, then restored the fixed bytes with matching
SHA-256 hashes.
- **Green:** 382 durable-chat, native-continuation, Stop and recovery
tests pass. The real GitHub/PostgreSQL handoff matrix passes 12/12. It
covers callback, application restart, callback/sweep race, valid input
behind a deleted source, and denial for unknown stop, wrong allocation,
paused endpoint, revoked principal, deleted or changed source, changed
conversation generation and foreign scope. Repeated callbacks preserve
one receipt, its actor, one successor and the unchanged stopped
predecessor.
- **Workspace continuation red/green:** the real PostgreSQL preparation
flow created a fresh attempt and saved a marker, then rejected the next
ordinary turn with the same binding mismatch observed in deployed use.
The repaired preparation and filesystem suites pass 86/86: stable and
attempt-root reuse, saved-marker preservation, a later clean sibling,
changed current bindings, foreign company/task, closed workspace,
traversal, missing path and symlink refusal. No task row reset or
old-turn replay is used.
- **Qualification red/green:** a native saved-agent qualification
returned 422, and its normal Test skipped qualification. The repaired
path passes 224 qualification/immutable-snapshot/environment tests and
176 UI/upgrade tests. It covers revoked authority, foreign or changed
computer, current-proof reuse, failed/missing proof, draft settings,
settings changed during the probe, and external/pool previews. On exact
current source, the real local Harness / Runtime UI also qualified a
newly saved disposable Computer with missing proof through actual
Daytona, then completed the actual OpenCode/OpenRouter hello probe using
`openrouter/deepseek/deepseek-v4.1-flash-1017`. Supported availability
changed from Legacy-only to Paperclip Runner after proof persistence.
Qualification and runtime leases both completed normal release/cleanup.
The five prior task run IDs were unchanged; no task wake was created.
This is macOS source-server/UI with real Linux Daytona and provider
services, not a hosted UI claim. Test probes are archived by the
existing inspection contract; release does not claim physical deletion.
- Full workspace `pnpm -r typecheck`, `pnpm build`, token, map and diff
checks pass at `f6450f4739fc8141815b8b7c35fde22a655c5f90`. The broad
local `pnpm test:run` passed 24,899 tests and reported one existing
nested-clone test timeout at 15 seconds. The complete failing file then
passed 21/21 with the normal timeout; the exact case also passed on the
pinned earlier baseline. The broad runner stopped before its later
groups. No timeout or production workaround was added. Prior head
`f6450f4` has 53 successful checks, two conditional Storybook skips and
Greptile 5/5 with no inline threads. New workspace-reuse head
`fcb13399ec70f92f71bd8250081251dca3f48962` passes workspace typecheck,
full build, 86 scoped tests, maps and diff checks; Current-head CI has
53 successful checks and two conditional Storybook skips; Greptile is
5/5 with no inline findings. Deployed ordinary-turn reuse acceptance
passed as described below.

- **Hosted Stop proof:** the real GitHub issue-mention path used native
OpenCode, `openrouter/deepseek/deepseek-v4.1-flash-1017`, Daytona and an
exact low-trust task boundary. Normal UI Stop interrupted an active Bash
call. Native and shared acknowledgements matched the exact lease.
Automatic recovery saved exactly `STARTED\n` (8 bytes), with no finish
marker. Reload showed Stopped and saved work recovered. Only the
original cancelled run remained. Normal five-minute cleanup completed
its exact provider-destruction transition.
- **Limits:** the provider buffered command output, so the marker was
read after recovery. The approved local provider credential could not
inspect the hosted sandbox; its 404 is not deletion proof. Cleanup
success comes from the exact lease's platform destruction transition. An
earlier hosted stale instruction disappeared after rollout with
unchanged raw run/action data; that historical cause remains
unexplained. Original incident holds and publication receipts were
untouched.
- **Hosted handoff proof:** after QA deployed the combined candidate
containing `46b2e99`, startup admitted the previously saved GitHub
author reply without another comment, Retry or manual action. The new
native run used the exact original wake receipt and saved comment. It
completed once, committed an external-response wait, preserved the
stopped predecessor, emptied the queue and left the task In review. The
answer appeared in the original GitHub conversation. Workspace
finalization and the exact lease's normal destruction transition
succeeded.
- **Hosted workspace proof:** on the combined candidate containing
`fcb13399`, normal Retry preserved the original authenticated GitHub
request and completed a fresh native attempt. One later ordinary
same-head author reply then completed in the exact same workspace ID and
attempt directory, with no Retry/fresh directive. Its input tree digest
matched the Retry's finalized digest; final copyback preserved that
digest. The persisted assessment row and all five prior run records were
unchanged, exactly one new reply run completed, and the task returned to
In review with an empty queue and no live run. Both exact leases
completed normal verified destruction. This used real native OpenCode,
the exact DeepSeek/OpenRouter model and Daytona, without workspace-row
repair or a new remote marker.
- **Publication limit:** the workspace audit compares the persisted
assessment and run history. The ordinary reply exposed a separate
destination-summary fallback that changed the visible GitHub summary
despite the preserved assessment row. The publication track is repairing
that behavior; this proof does not claim the visible summary was
preserved.
- **Deployment boundary:** current-head CI/review and hosted same-source
Retry/ordinary-workspace reuse acceptance passed. The separate
visible-summary publication failure remains owned by the publication
track. Combined hosted UI confirmation of the native-agent Test path is
coordinated separately. Actual missing-proof UI/provider qualification
has passed in the isolated local instance above; prior hosted Stop and
same-input handoff proof remains separate. No merge or deployment is
performed by this pull request.

## Risks

Recovery state and termination receipts must match the exact source run
and allocation. Missing or foreign proof keeps the hold. The chat owner
must retain the original receipt and actor and recheck current
authority. Cleanup cannot grant authority or replay the stopped turn.
Existing native ownership, Stop, budget, permission and uncertain-effect
guards remain in place.

## Model Used

OpenAI GPT-6.1 Sol (`gpt-6.1-sol`) through Codex, with reasoning,
repository tools, source editing and test execution. The context-window
size is not exposed in this session.

## Checklist

- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass (affected suites and
isolated failing file pass; broad timeout retained above)
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green (current fcb has 53 successful
checks/two conditional Storybook skips)
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge

---------

Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-10-26 22:21:58 -05:00

1451 lines
52 KiB
TypeScript

import { EMBEDDED_POSTGRES_TEST_TIMEOUT_MS } from "@paperclipai/db";
import { randomUUID } from "node:crypto";
import { and, eq } from "drizzle-orm";
import {
afterAll,
afterEach,
beforeAll,
describe,
expect,
it,
vi,
} from "vitest";
import {
agents,
companies,
authUsers,
companyMemberships,
createDb,
heartbeatRuns,
agentWakeupRequests,
chatConversations,
chatActions,
chatDeliveries,
chatEndpoints,
chatMessageLinks,
chatPublications,
issueComments,
issueRecoveryActions,
issues,
toolApplications,
toolConnections,
} from "@paperclipai/db";
import { startEmbeddedPostgresTestDatabase } from "./helpers/embedded-postgres.js";
import { heartbeatService } from "../services/heartbeat.js";
import {
authorizeFailedChatRunRetryWake,
authorizeCommittedChatResponse,
CommittedChatResponseAuthorizationError,
registerCommittedChatResponseAuthority,
createDurableChatWakeupRequest,
FailedChatRunRetryAuthorizationError,
registerFailedChatRunRetryAuthority,
} from "../services/durable-chat-wakeup.js";
import { conflict } from "../errors.js";
import type { ServerAdapterModule } from "../adapters/index.js";
import { queueIssueAssignmentWakeup } from "../services/issue-assignment-wakeup.js";
import {
registerServerAdapter,
runningProcesses,
unregisterServerAdapter,
} from "../adapters/index.js";
describe("durable inbound chat scheduler receipts", () => {
let temporary: Awaited<ReturnType<typeof startEmbeddedPostgresTestDatabase>>;
let db: ReturnType<typeof createDb>;
const liveRunIds = new Set<string>();
const unregisterAuthorities: Array<() => void> = [];
const execute = vi.fn<ServerAdapterModule["execute"]>(async (input) => {
const issueId = String(input.context.issueId);
await db.insert(issueComments).values({
companyId: input.agent.companyId,
issueId,
authorAgentId: input.agent.id,
createdByRunId: input.runId,
body: "Exact retry dispatched",
});
await db
.update(issues)
.set({ status: "done" })
.where(eq(issues.id, issueId));
return {
exitCode: 0,
signal: null,
timedOut: false,
summary: "Exact retry dispatched",
};
});
beforeAll(async () => {
temporary = await startEmbeddedPostgresTestDatabase(
"chat-wakeup-receipts-",
);
db = createDb(temporary.connectionString);
registerServerAdapter({
type: "durable_chat_retry_test",
execute,
testEnvironment: async () => ({
adapterType: "durable_chat_retry_test",
status: "pass",
checks: [],
testedAt: new Date(0).toISOString(),
}),
});
}, EMBEDDED_POSTGRES_TEST_TIMEOUT_MS);
afterEach(() => {
execute.mockClear();
for (const unregister of unregisterAuthorities.splice(0)) unregister();
for (const runId of liveRunIds) runningProcesses.delete(runId);
liveRunIds.clear();
});
afterAll(async () => {
unregisterServerAdapter("durable_chat_retry_test");
await temporary.cleanup();
});
async function fixture(deferred = false) {
const companyId = randomUUID(),
agentId = randomUUID(),
issueId = randomUUID(),
activeRunId = randomUUID();
await db.insert(companies).values({
id: companyId,
name: "Durable wake",
issuePrefix: `D${companyId.slice(0, 7)}`,
defaultResponsibleUserId: "board-user",
requireBoardApprovalForNewAgents: false,
});
await db.insert(authUsers).values({ id: "board-user", name: "Fixture user", email: "board-user@example.test", createdAt: new Date(), updatedAt: new Date() }).onConflictDoNothing();
await db.insert(companyMemberships).values({ companyId, principalType: "user", principalId: "board-user", status: "active", membershipRole: "member" });
await db.insert(agents).values({
id: agentId,
companyId,
name: "Maya",
role: "ceo",
status: "running",
adapterType: "process",
adapterConfig: {},
runtimeConfig: { heartbeat: { maxConcurrentRuns: 1 } },
});
await db.insert(heartbeatRuns).values({
id: activeRunId,
companyId,
agentId,
status: "running",
// This fixture-owned slot has no provider process/work. Its retirement
// must not impersonate an ambiguous legacy provider failure.
resultJson: { executionRecovery: { kind: "bootstrap", providerWorkStarted: false } },
contextSnapshot: { issueId: deferred ? issueId : randomUUID() },
});
runningProcesses.set(activeRunId, {
child: {} as never,
graceSec: 0,
processGroupId: null,
});
liveRunIds.add(activeRunId);
await db.insert(issues).values({
id: issueId,
companyId,
title: "Bound chat task",
status: "in_progress",
assigneeAgentId: agentId,
responsibleUserId: "board-user",
...(deferred
? {
executionRunId: activeRunId,
executionAgentNameKey: "maya",
executionLockedAt: new Date(),
}
: {}),
});
const heartbeat = heartbeatService(db);
const authorize = vi.fn(async () => {});
const request = (commentId = randomUUID(), actor = "board-user") =>
createDurableChatWakeupRequest({
id: randomUUID(),
companyId,
agentId,
issueId,
commentId,
requestedByActorType: "user",
requestedByActorId: actor,
requestedAt: new Date(),
authorize,
});
const wake = (durableChatRequest: ReturnType<typeof request>) =>
queueIssueAssignmentWakeup({
heartbeat,
issue: { id: issueId, assigneeAgentId: agentId, status: "in_progress" },
reason: "External chat message received",
mutation: "chat_message_received",
contextSource: "chat:slack",
requestedByActorType: "user",
requestedByActorId: durableChatRequest.requestedByActorId,
wakeCommentId: durableChatRequest.commentId,
durableChatRequest,
rethrowOnError: true,
});
return {
companyId,
agentId,
issueId,
activeRunId,
heartbeat,
authorize,
request,
wake,
};
}
it.each([
...["unmentioned", "mentioned", "coalesced", "recovery", "deferred"].map((scenario) => ({ scenario, isDM: false })),
...["unmentioned", "coalesced"].map((scenario) => ({ scenario, isDM: true })),
])("checks Slack channel mentions after a real heartbeat queue wait ($scenario, DM=$isDM)", async ({ scenario, isDM }) => {
const mentioned = scenario === "mentioned" || scenario === "coalesced";
const allowed = isDM || scenario === "mentioned" || scenario === "recovery";
const f = await fixture();
await db.update(agents).set({ adapterType: "durable_chat_retry_test" }).where(eq(agents.id, f.agentId));
const applicationId = randomUUID(), connectionId = randomUUID(), endpointId = randomUUID();
await db.insert(toolApplications).values({ id: applicationId, companyId: f.companyId, name: "Mention gate", type: "chat" });
await db.insert(toolConnections).values({ id: connectionId, companyId: f.companyId, applicationId,
name: "Mention gate", uid: randomUUID(), connectionPurpose: "channel", transport: "chat_sdk", status: "active" });
await db.insert(chatEndpoints).values({ id: endpointId, companyId: f.companyId, connectionId,
provider: "slack", publicId: randomUUID(), assignedAgentId: f.agentId, status: "active" });
const request = f.request();
await db.insert(issueComments).values({ id: request.commentId, companyId: f.companyId, issueId: f.issueId,
authorUserId: "board-user", body: mentioned ? "@bot work" : "Work" });
const [delivery] = await db.insert(chatDeliveries).values({ companyId: f.companyId, endpointId,
providerEventId: randomUUID(), deduplicationKey: randomUUID(), eventKind: isDM ? "direct_message" : "message", state: "processed",
normalizedEvent: { trigger: isDM ? "direct_message" : "subscribed_message", conversation: { isDirectMessage: isDM }, message: { mentionedBot: mentioned } } }).returning();
await db.insert(chatActions).values({ id: request.id, companyId: f.companyId, endpointId,
deliveryId: delivery.id, kind: "inbound_wakeup", providerActionId: randomUUID(), status: "processed",
payload: { issueId: f.issueId, commentId: request.commentId } });
await f.wake(request);
const [receipt] = await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.id, request.id));
expect(receipt.status).toBe("queued");
const [queuedRun] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, receipt.runId!));
expect(queuedRun.startedAt).toBeNull();
const originalStartedAt = new Date(Date.now() - 60_000);
if (scenario === "recovery") {
// Same-run retries retain their first claim time, even when queued again.
await db.update(heartbeatRuns).set({ startedAt: originalStartedAt })
.where(eq(heartbeatRuns.id, receipt.runId!));
}
let deferredRequestId: string | undefined;
if (scenario === "coalesced" || scenario === "deferred") {
const second = f.request();
await db.insert(issueComments).values({ id: second.commentId, companyId: f.companyId, issueId: f.issueId,
authorUserId: "board-user", body: scenario === "deferred" ? "@bot follow-up" : "Unmentioned follow-up" });
const [secondDelivery] = await db.insert(chatDeliveries).values({ companyId: f.companyId, endpointId,
providerEventId: randomUUID(), deduplicationKey: randomUUID(), eventKind: isDM ? "direct_message" : "message", state: "processed",
normalizedEvent: { trigger: isDM ? "direct_message" : "subscribed_message", conversation: { isDirectMessage: isDM }, message: { mentionedBot: scenario === "deferred" } } }).returning();
await db.insert(chatActions).values({ id: second.id, companyId: f.companyId, endpointId,
deliveryId: secondDelivery.id, kind: "inbound_wakeup", providerActionId: randomUUID(), status: "processed",
payload: { issueId: f.issueId, commentId: second.commentId } });
await db.update(issues).set({ executionRunId: receipt.runId, executionAgentNameKey: "maya", executionLockedAt: new Date() })
.where(eq(issues.id, f.issueId));
if (scenario === "deferred") {
// Seed the persisted queue state left by a deferred follow-up before
// its predecessor is returned to the unstarted queue.
await db.update(heartbeatRuns).set({ status: "running" }).where(eq(heartbeatRuns.id, receipt.runId!));
runningProcesses.set(receipt.runId!, { child: {} as never, graceSec: 0, processGroupId: null });
liveRunIds.add(receipt.runId!);
}
await f.wake(second);
const [coalesced] = await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.id, second.id));
if (scenario === "deferred") {
expect(coalesced).toMatchObject({ status: "deferred_issue_execution", runId: null });
deferredRequestId = second.id;
runningProcesses.delete(receipt.runId!);
liveRunIds.delete(receipt.runId!);
await db.update(heartbeatRuns).set({ status: "queued" }).where(eq(heartbeatRuns.id, receipt.runId!));
} else {
expect(coalesced).toMatchObject({ status: "coalesced", runId: receipt.runId });
}
}
await db.update(chatEndpoints).set({ requireAtMention: true }).where(eq(chatEndpoints.id, endpointId));
await f.heartbeat.cancelRun(f.activeRunId, "Fixture slot released", { suppressImmediateRecovery: true });
await f.heartbeat.drainActiveRunExecutions();
const [run] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, receipt.runId!));
expect(run.status).toBe(allowed ? "succeeded" : "cancelled");
expect(execute).toHaveBeenCalledTimes(allowed || scenario === "deferred" ? 1 : 0);
if (deferredRequestId) {
const [promoted] = await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.id, deferredRequestId));
const [followup] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, promoted.runId!));
expect(followup.status).toBe("succeeded");
}
if (!allowed) expect(run.errorCode).toBe("chat_mention_required");
if (scenario === "recovery") expect(run.startedAt).toEqual(originalStartedAt);
});
async function executionHold(f: { companyId: string; issueId: string; agentId: string }) {
const [action] = await db.insert(issueRecoveryActions).values({
companyId: f.companyId, sourceIssueId: f.issueId, kind: "active_run_watchdog",
ownerType: "board", returnOwnerAgentId: f.agentId,
cause: "legacy_execution_requires_reconciliation", status: "resolved",
fingerprint: randomUUID(), evidence: { automaticRecovery: { replay: "blocked" } },
nextAction: "Check the stopped execution.",
}).returning();
return () => db.update(issueRecoveryActions).set({ evidence: {} }).where(eq(issueRecoveryActions.id, action!.id));
}
it("defers held inbound chat exactly once and keeps its authority separate from a generic wake", async () => {
const f = await fixture();
const clearHold = await executionHold(f);
const request = f.request();
const wake = () => f.heartbeat.wakeup(f.agentId, {
source: "on_demand", triggerDetail: "manual", reason: "issue_commented",
payload: { issueId: f.issueId, commentId: request.commentId },
contextSnapshot: { issueId: f.issueId, source: "chat:slack", wakeCommentId: request.commentId },
requestedByActorType: "user", requestedByActorId: request.requestedByActorId,
durableChatRequest: request,
});
for (let i = 0; i < 3; i++) expect(await wake()).toBeNull();
const receipts = await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.agentId, f.agentId));
expect(receipts).toHaveLength(1);
expect(receipts[0]).toMatchObject({
id: request.id, idempotencyKey: request.idempotencyKey, requestedAt: request.requestedAt,
status: "deferred_issue_execution", runId: null, requestedByActorId: request.requestedByActorId,
payload: { _paperclipWakeContext: { source: "chat:slack", wakeCommentIds: [request.commentId] } },
});
expect(f.authorize).toHaveBeenCalledTimes(1);
expect(await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.wakeupRequestId, request.id))).toHaveLength(0);
await clearHold();
const generic = await f.heartbeat.wakeup(f.agentId, {
source: "on_demand", triggerDetail: "manual", reason: "issue_resumed",
payload: { issueId: f.issueId }, requestedByActorType: "user", requestedByActorId: "board-user",
});
expect(generic?.status).toBe("queued");
expect(generic?.contextSnapshot?.wakeCommentIds).toBeUndefined();
expect((await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.id, request.id)))[0]).toMatchObject({
status: "deferred_issue_execution", runId: null,
});
});
it("reconsiders the same held inbound receipt only through its live durable owner", async () => {
const f = await fixture();
const clearHold = await executionHold(f);
const request = f.request();
expect(await f.wake(request)).toBeNull();
await clearHold();
const reconsidered = createDurableChatWakeupRequest({ ...request, resumeDeferred: true });
const [first, second] = await Promise.all([f.wake(reconsidered), f.wake(reconsidered)]);
const receipts = await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.agentId, f.agentId));
expect(receipts).toHaveLength(1);
expect(receipts[0]).toMatchObject({ id: request.id, status: "queued", requestedAt: request.requestedAt });
expect(await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.wakeupRequestId, request.id))).toHaveLength(1);
expect(first?.id ?? second?.id).toBe(receipts[0].runId);
expect(f.authorize.mock.calls.length).toBeGreaterThan(1);
});
async function retryFixture(deferred = false) {
const f = await fixture(deferred);
const applicationId = randomUUID(),
connectionId = randomUUID(),
endpointId = randomUUID();
const failedRunId = randomUUID(),
actionId = randomUUID();
const commentIds = [randomUUID(), randomUUID()];
const taskKey = `D${f.companyId.slice(0, 7)}-1`;
await db
.update(issues)
.set({ identifier: taskKey })
.where(eq(issues.id, f.issueId));
await db.insert(toolApplications).values({
id: applicationId,
companyId: f.companyId,
name: "Retry authority",
type: "chat",
});
await db.insert(toolConnections).values({
id: connectionId,
companyId: f.companyId,
applicationId,
name: "Retry",
uid: `retry-${connectionId}`,
connectionPurpose: "channel",
transport: "chat_sdk",
status: "active",
});
await db.insert(chatEndpoints).values({
id: endpointId,
companyId: f.companyId,
connectionId,
provider: "slack",
publicId: randomUUID(),
assignedAgentId: f.agentId,
status: "active",
});
await db.insert(heartbeatRuns).values({
id: failedRunId,
companyId: f.companyId,
agentId: f.agentId,
status: "failed",
finishedAt: new Date(),
contextSnapshot: {
issueId: f.issueId,
taskKey,
source: "chat:slack",
wakeCommentIds: commentIds,
},
});
await db.insert(issueComments).values(
commentIds.map((id) => ({
id,
companyId: f.companyId,
issueId: f.issueId,
authorUserId: "board-user",
body: "Exact original request",
})),
);
await db.insert(chatActions).values({
id: actionId,
companyId: f.companyId,
endpointId,
kind: "failed_run_retry",
providerActionId: `failed_run_retry:${failedRunId}`,
status: "issued",
payload: { issueId: f.issueId, failedRunId },
});
const context = {
issueId: f.issueId,
source: "chat:slack",
taskKey,
wakeCommentId: commentIds[1],
wakeCommentIds: commentIds,
retryOfRunId: failedRunId,
chatFailedRunRetry: { version: 1, actionId, failedRunId },
};
const request = createDurableChatWakeupRequest({
...f.request(commentIds[1]),
id: actionId,
failedRunRetry: { failedRunId },
});
const authority = vi.fn(
async (
_tx: unknown,
input: { contextSnapshot: Record<string, unknown>; phase: string },
) => {
expect(input.contextSnapshot).toMatchObject(context);
},
);
const register = () => {
const dispose = registerFailedChatRunRetryAuthority(db, authority);
unregisterAuthorities.push(dispose);
return dispose;
};
const wake = (options: Record<string, unknown> = {}) =>
f.heartbeat.wakeup(f.agentId, {
source: "assignment",
triggerDetail: "system",
reason: "retry_failed_chat_run",
payload: { ...context },
contextSnapshot: { ...context },
requestedByActorType: "user",
requestedByActorId: "board-user",
durableChatRequest: request,
allowRunCoalescing: true,
...options,
});
return {
...f,
failedRunId,
actionId,
context,
commentIds,
request,
authority,
register,
wake,
};
}
it("fails closed for a persisted retry when its live authority is unavailable", async () => {
const f = await retryFixture();
await expect(f.wake()).rejects.toBeInstanceOf(
FailedChatRunRetryAuthorizationError,
);
expect(
await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.id, f.actionId)),
).toEqual([]);
});
it("does not turn a forged retry selector into authority", async () => {
const f = await fixture();
await expect(
f.heartbeat.wakeup(f.agentId, {
payload: { issueId: f.issueId },
contextSnapshot: {
chatFailedRunRetry: {
version: 1,
actionId: randomUUID(),
failedRunId: randomUUID(),
},
},
requestedByActorType: "user",
requestedByActorId: "board-user",
}),
).rejects.toBeInstanceOf(FailedChatRunRetryAuthorizationError);
expect(
await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.agentId, f.agentId)),
).toEqual([]);
});
it("preserves exact retry provenance and a deeply frozen server-only selector", async () => {
const f = await retryFixture();
f.register();
expect(Object.isFrozen(f.request.failedRunRetry)).toBe(true);
const run = await f.wake();
expect(run).toMatchObject({
status: "queued",
retryOfRunId: f.failedRunId,
wakeupRequestId: f.actionId,
contextSnapshot: f.context,
});
expect(f.authority).toHaveBeenCalledTimes(1);
expect(f.authority.mock.calls[0][1]).toMatchObject({ phase: "admission" });
await f.wake();
expect(f.authority).toHaveBeenCalledTimes(1);
});
it("cannot coalesce an exact retry into active or deferred ordinary work", async () => {
const f = await retryFixture(true);
const ordinary = f.request;
await queueIssueAssignmentWakeup({
heartbeat: f.heartbeat,
issue: {
id: f.issueId,
assigneeAgentId: f.agentId,
status: "in_progress",
},
reason: "External chat message received",
mutation: "chat_message_received",
contextSource: "chat:slack",
requestedByActorType: "user",
requestedByActorId: "board-user",
wakeCommentId: f.commentIds[0],
durableChatRequest: createDurableChatWakeupRequest({
...ordinary,
id: randomUUID(),
commentId: f.commentIds[0],
failedRunRetry: undefined,
}),
rethrowOnError: true,
});
f.register();
await f.wake();
const receipts = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.agentId, f.agentId));
expect(receipts).toHaveLength(2);
expect(
receipts.every(
(row) =>
row.status === "deferred_issue_execution" && row.coalescedCount === 0,
),
).toBe(true);
expect(
await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.wakeupRequestId, f.actionId)),
).toEqual([]);
});
it("rediscovers a retry from durable provenance even if its hint was stripped", async () => {
const f = await retryFixture();
const { chatFailedRunRetry: _selector, ...contextSnapshot } = f.context;
await expect(
authorizeFailedChatRunRetryWake(db, db, {
phase: "execution",
wakeupRequestId: f.actionId,
companyId: f.companyId,
agentId: f.agentId,
issueId: f.issueId,
contextSnapshot,
}),
).rejects.toBeInstanceOf(FailedChatRunRetryAuthorizationError);
});
it("an older service disposer cannot remove a newer retry authority", async () => {
const f = await retryFixture();
const oldDispose = f.register();
const replacement = vi.fn(async () => {});
unregisterAuthorities.push(
registerFailedChatRunRetryAuthority(db, replacement),
);
oldDispose();
await f.wake();
expect(replacement).toHaveBeenCalledTimes(1);
expect(f.authority).not.toHaveBeenCalled();
});
it.each([false, true])("cancels a revoked deferred retry without reopening its batch (execution hold=%s)", async (held) => {
const f = await retryFixture(true);
const clearHold = held ? await executionHold(f) : null;
f.register();
await f.wake();
await clearHold?.();
f.authority.mockImplementation(async (_tx, input) => {
if (input.phase === "promotion")
throw conflict("Current chat access was revoked");
});
await db
.update(issues)
.set({ status: "done" })
.where(eq(issues.id, f.issueId));
await f.heartbeat.cancelRun(f.activeRunId, "Fixture predecessor finished", {
suppressImmediateRecovery: true,
});
await f.heartbeat.drainActiveRunExecutions();
const [receipt] = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.id, f.actionId));
expect(receipt).toMatchObject({ status: "cancelled", runId: null });
expect(receipt.payload).toMatchObject({ wakeCommentIds: f.commentIds });
const [issue] = await db
.select()
.from(issues)
.where(eq(issues.id, f.issueId));
expect(issue).toMatchObject({ status: "done", executionRunId: null });
expect(
await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.wakeupRequestId, f.actionId)),
).toEqual([]);
expect(f.authority.mock.calls.map((call) => call[1].phase)).toEqual([
"admission",
"promotion",
]);
});
it("denies a queued retry at execution when current authority is revoked", async () => {
const f = await retryFixture();
f.register();
const queued = await f.wake();
f.authority.mockImplementation(async (_tx, input) => {
if (input.phase === "execution")
throw conflict("Current chat access was revoked", {
code: "chat_failed_run_retry_not_authorized",
});
});
await f.heartbeat.cancelRun(
f.activeRunId,
"Fixture execution slot released",
{ suppressImmediateRecovery: true },
);
await f.heartbeat.drainActiveRunExecutions();
const [run] = await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.id, queued!.id));
expect(run).toMatchObject({
status: "failed",
retryOfRunId: f.failedRunId,
nativeSessionId: null,
errorCode: "chat_failed_run_retry_not_authorized",
});
expect(run.error).toContain("Current chat access was revoked");
const [issue] = await db
.select()
.from(issues)
.where(eq(issues.id, f.issueId));
expect(issue).toMatchObject({ status: "blocked", executionRunId: null });
expect(
await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.agentId, f.agentId)),
).toHaveLength(3);
expect(f.authority.mock.calls.map((call) => call[1].phase)).toEqual([
"admission",
"execution",
]);
});
it.each([false, true])("preserves the exact retry column and comment batch during promotion (execution hold=%s)", async (held) => {
const f = await retryFixture(true);
const clearHold = held ? await executionHold(f) : null;
f.register();
await f.wake();
await clearHold?.();
// A separate fixture-owned run occupies the agent slot after the issue's
// predecessor releases it, so this asserts promotion before execution.
const blockerId = randomUUID();
await db.insert(heartbeatRuns).values({
id: blockerId,
companyId: f.companyId,
agentId: f.agentId,
status: "running",
contextSnapshot: { issueId: randomUUID() },
});
runningProcesses.set(blockerId, {
child: {} as never,
graceSec: 0,
processGroupId: null,
});
liveRunIds.add(blockerId);
await f.heartbeat.cancelRun(f.activeRunId, "Fixture predecessor finished", {
suppressImmediateRecovery: true,
});
await f.heartbeat.drainActiveRunExecutions();
const [run] = await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.wakeupRequestId, f.actionId));
expect(run).toMatchObject({
status: "queued",
retryOfRunId: f.failedRunId,
contextSnapshot: f.context,
});
const [receipt] = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.id, f.actionId));
expect(receipt).toMatchObject({ status: "queued", runId: run.id });
expect(f.authority.mock.calls.map((call) => call[1].phase)).toEqual([
"admission",
"promotion",
]);
});
it.each([false, true])(
"rechecks current retry authority immediately before dispatch (revoked=%s)",
async (revoked) => {
const f = await retryFixture();
await db
.update(agents)
.set({ adapterType: "durable_chat_retry_test" })
.where(eq(agents.id, f.agentId));
f.register();
const queued = await f.wake();
let executionChecks = 0;
f.authority.mockImplementation(async (_tx, input) => {
if (input.phase === "execution" && ++executionChecks === 2 && revoked) {
throw conflict("Current chat access was revoked before dispatch", {
code: "chat_failed_run_retry_not_authorized",
});
}
});
await f.heartbeat.cancelRun(
f.activeRunId,
"Fixture execution slot released",
{ suppressImmediateRecovery: true },
);
await f.heartbeat.drainActiveRunExecutions();
const [run] = await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.id, queued!.id));
expect(executionChecks).toBe(2);
expect(execute).toHaveBeenCalledTimes(revoked ? 0 : 1);
expect(run).toMatchObject({
status: revoked ? "failed" : "succeeded",
retryOfRunId: f.failedRunId,
nativeSessionId: null,
});
if (revoked)
expect(run.errorCode).toBe("chat_failed_run_retry_not_authorized");
expect(
await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.agentId, f.agentId)),
).toHaveLength(3);
expect(f.authority.mock.calls.map((call) => call[1].phase)).toEqual([
"admission",
"execution",
"execution",
]);
},
);
it.each([
{ forceFreshSession: true },
{ resumeFromRunId: randomUUID() },
{ resumeSessionParams: { sessionId: "unrelated-provider-session" } },
{ resumeSessionDisplayId: "unrelated-provider-session" },
])(
"does not admit routing overrides on an exact retry (%j)",
async (override) => {
const f = await retryFixture();
f.register();
await expect(
f.wake({ contextSnapshot: { ...f.context, ...override } }),
).rejects.toBeInstanceOf(FailedChatRunRetryAuthorizationError);
expect(f.authority).not.toHaveBeenCalled();
expect(
await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.id, f.actionId)),
).toEqual([]);
},
);
it.each([
{ withDeferredInput: false, sourceKind: "inbound_wakeup" },
{ withDeferredInput: true, sourceKind: "inbound_wakeup" },
{ withDeferredInput: false, sourceKind: "unrelated_action" },
{ withDeferredInput: false, sourceKind: "missing" },
].flatMap((scope) => [false, true].map((preProvider) => ({ ...scope, preProvider }))))(
"fences generic recovery by actual original chat ownership ($sourceKind, deferred=$withDeferredInput, pre-provider=$preProvider)",
async ({ withDeferredInput, sourceKind, preProvider }) => {
const f = await retryFixture();
await db
.update(agents)
.set({ adapterType: "durable_chat_retry_test" })
.where(eq(agents.id, f.agentId));
await db
.update(chatActions)
.set({ kind: sourceKind, payload: { issueId: f.issueId } })
.where(eq(chatActions.id, f.actionId));
if (sourceKind === "missing")
await db.delete(chatActions).where(eq(chatActions.id, f.actionId));
const genericRecoveryExpected = preProvider && sourceKind !== "inbound_wakeup";
const nextExecutionExpected = preProvider && (withDeferredInput || genericRecoveryExpected);
const { failedRunRetry: _retry, ...ordinaryRequest } = f.request;
const wakeOriginal = (id: string, commentId: string) =>
f.heartbeat.wakeup(f.agentId, {
source: "assignment",
triggerDetail: "system",
reason: "chat_message_received",
requestedByActorType: "user",
requestedByActorId: "board-user",
payload: {
issueId: f.issueId,
wakeCommentId: commentId,
wakeCommentIds: [commentId],
},
contextSnapshot: {
issueId: f.issueId,
taskKey: f.context.taskKey,
source: "chat:slack",
wakeCommentId: commentId,
wakeCommentIds: [commentId],
},
durableChatRequest: createDurableChatWakeupRequest({
...ordinaryRequest,
id,
commentId,
}),
allowRunCoalescing: false,
});
if (preProvider) {
execute.mockResolvedValueOnce({
exitCode: 1, signal: null, timedOut: false,
summary: "Fixture failed before provider work",
resultJson: { executionRecovery: { kind: "bootstrap", providerWorkStarted: false } },
});
} else {
execute.mockRejectedValueOnce(new Error("Original chat provider outcome is unknown"));
}
const original = await wakeOriginal(f.actionId, f.commentIds[1]);
const deferredId = randomUUID();
if (withDeferredInput) {
const [owner] = await db
.select()
.from(chatActions)
.where(eq(chatActions.id, f.actionId));
await db.insert(chatActions).values({
...owner,
id: deferredId,
providerActionId: `wakeup:${deferredId}`,
});
expect(await wakeOriginal(deferredId, f.commentIds[0])).toBeNull();
const [receipt] = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.id, deferredId));
expect(receipt.status).toBe("deferred_issue_execution");
}
await f.heartbeat.cancelRun(
f.activeRunId,
"Fixture execution slot released",
{ suppressImmediateRecovery: true },
);
await f.heartbeat.drainActiveRunExecutions();
const [run] = await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.id, original!.id));
expect(run).toMatchObject({
status: "failed",
retryOfRunId: null,
wakeupRequestId: f.actionId,
});
expect(execute).toHaveBeenCalledTimes(
nextExecutionExpected ? 2 : 1,
);
const runs = await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.agentId, f.agentId));
expect(runs).toHaveLength(
nextExecutionExpected ? 4 : 3,
);
expect(
runs.some(
(entry) =>
(entry.contextSnapshot as Record<string, unknown>)?.source ===
"issue.continuation_recovery",
),
).toBe(genericRecoveryExpected);
const [issue] = await db
.select()
.from(issues)
.where(eq(issues.id, f.issueId));
expect(issue).toMatchObject({
status:
nextExecutionExpected ? "done" : preProvider ? "blocked" : "in_progress",
executionRunId: null,
});
if (withDeferredInput) {
const [receipt] = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.id, deferredId));
const next = runs.find((entry) => entry.id === receipt.runId);
if (!preProvider) {
// Unknown earlier effects require reconciliation. Preserve the
// separate user's durable cause, without dispatching it on this
// cancellation/failure path or relabeling it as an automatic retry.
expect(receipt).toMatchObject({ status: "deferred_issue_execution", runId: null });
expect(next).toBeUndefined();
return;
}
expect(next).toMatchObject({
status: "succeeded",
retryOfRunId: null,
wakeupRequestId: deferredId,
contextSnapshot: {
source: "chat:slack",
wakeCommentIds: [f.commentIds[0]],
},
});
}
},
);
it("keeps committed-response authority exact-Db scoped and separate from retry hints", async () => {
const input = {
companyId: randomUUID(),
issueId: randomUUID(),
agentId: randomUUID(),
runId: randomUUID(),
resultId: randomUUID(),
};
const otherDb = {} as typeof db;
const authority = vi.fn(async () => undefined);
const dispose = registerCommittedChatResponseAuthority(db, authority);
try {
await expect(
authorizeCommittedChatResponse(otherDb, db, input),
).rejects.toBeInstanceOf(CommittedChatResponseAuthorizationError);
expect(authority).not.toHaveBeenCalled();
await authorizeCommittedChatResponse(db, db, input);
expect(authority).toHaveBeenCalledExactlyOnceWith(db, input);
} finally {
dispose();
}
await expect(
authorizeCommittedChatResponse(db, db, input),
).rejects.toBeInstanceOf(CommittedChatResponseAuthorizationError);
});
it("preserves the current committed-response authority and exact transient failure", async () => {
const input = {
companyId: randomUUID(),
issueId: randomUUID(),
agentId: randomUUID(),
runId: randomUUID(),
resultId: randomUUID(),
};
const first = registerCommittedChatResponseAuthority(
db,
async () => undefined,
);
const storageFailure = new Error("temporary source storage failure");
const second = registerCommittedChatResponseAuthority(db, async () => {
throw storageFailure;
});
try {
first();
await expect(authorizeCommittedChatResponse(db, db, input)).rejects.toBe(
storageFailure,
);
} finally {
second();
}
});
it("does not discard a deferred retry on transient authority storage failure", async () => {
const f = await retryFixture(true);
f.register();
await f.wake();
const failure = new Error("Fixture authorization store unavailable");
f.authority.mockRejectedValueOnce(failure);
await expect(
f.heartbeat.cancelRun(f.activeRunId, "Fixture predecessor finished", {
suppressImmediateRecovery: true,
}),
).rejects.toBe(failure);
const [receipt] = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.id, f.actionId));
expect(receipt).toMatchObject({
status: "deferred_issue_execution",
runId: null,
});
expect(receipt.payload).toMatchObject({ wakeCommentIds: f.commentIds });
expect(
await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.wakeupRequestId, f.actionId)),
).toEqual([]);
});
it("retries and competing workers create one queued receipt and one run", async () => {
const f = await fixture();
const request = f.request();
await Promise.all([f.wake(request), f.wake(request)]);
await f.wake(request);
const receipts = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.id, request.id));
expect(receipts).toHaveLength(1);
expect(receipts[0]).toMatchObject({
status: "queued",
idempotencyKey: request.idempotencyKey,
});
expect(
await db
.select()
.from(heartbeatRuns)
.where(
and(
eq(heartbeatRuns.agentId, f.agentId),
eq(heartbeatRuns.wakeupRequestId, request.id),
),
),
).toHaveLength(1);
expect(f.authorize).toHaveBeenCalledTimes(1);
});
it("records a stable receipt when merging into a deferred wake and never merges the replay twice", async () => {
const f = await fixture(true);
const first = f.request(),
second = f.request();
await f.wake(first);
await f.wake(second);
await f.wake(second);
const rows = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.agentId, f.agentId));
expect(rows).toHaveLength(2);
expect(rows.find((row) => row.id === first.id)).toMatchObject({
status: "deferred_issue_execution",
coalescedCount: 1,
});
expect(rows.find((row) => row.id === second.id)).toMatchObject({
status: "coalesced",
payload: { coalescedIntoWakeupRequestId: first.id },
});
expect(
await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.agentId, f.agentId)),
).toHaveLength(1);
});
it("keeps different actors in separate deferred receipts", async () => {
const f = await fixture(true);
await f.wake(f.request());
await f.wake(f.request(randomUUID(), "another-user"));
const rows = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.agentId, f.agentId));
expect(rows).toHaveLength(2);
expect(
rows.every(
(row) =>
row.status === "deferred_issue_execution" && row.coalescedCount === 0,
),
).toBe(true);
});
it("records queued-run coalescence once with the stable receipt ID", async () => {
const f = await fixture();
const first = f.request(),
second = f.request();
await f.wake(first);
const [firstReceipt] = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.id, first.id));
await db
.update(issues)
.set({
executionRunId: firstReceipt.runId,
executionAgentNameKey: "maya",
executionLockedAt: new Date(),
})
.where(eq(issues.id, f.issueId));
await f.wake(second);
await f.wake(second);
const [receipt] = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.id, second.id));
expect(receipt).toMatchObject({
status: "coalesced",
runId: firstReceipt.runId,
coalescedCount: 1,
});
expect(
await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.agentId, f.agentId)),
).toHaveLength(2);
});
it("rejects a stable receipt ID reused for a different authenticated actor", async () => {
const f = await fixture();
const first = f.request();
await f.wake(first);
const conflicting = createDurableChatWakeupRequest({
...first,
requestedByActorId: "different-user",
});
await expect(f.wake(conflicting)).rejects.toThrow(
"chat_inbound_wakeup_receipt_conflict",
);
expect(
await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.agentId, f.agentId)),
).toHaveLength(1);
expect(
await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.agentId, f.agentId)),
).toHaveLength(2);
});
it("never readmits or dispatches a cancelled receipt on replay", async () => {
const f = await fixture();
const request = f.request();
await f.wake(request);
const [receipt] = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.id, request.id));
await db
.update(agentWakeupRequests)
.set({ status: "cancelled" })
.where(eq(agentWakeupRequests.id, request.id));
await db
.update(heartbeatRuns)
.set({ status: "cancelled", finishedAt: new Date() })
.where(eq(heartbeatRuns.id, receipt.runId!));
await f.wake(request);
expect(f.authorize).toHaveBeenCalledTimes(1);
expect(
await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.agentId, f.agentId)),
).toHaveLength(2);
});
it("does not accept a JSON copy of the internal scheduling capability", async () => {
const f = await fixture();
const request = f.request();
await expect(f.wake(JSON.parse(JSON.stringify(request)))).rejects.toThrow(
"chat_inbound_wakeup_binding_denied",
);
expect(
await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.agentId, f.agentId)),
).toEqual([]);
});
it("rechecks current authorization in the scheduling transaction", async () => {
const f = await fixture();
f.authorize.mockRejectedValueOnce(new Error("reach revoked"));
await expect(f.wake(f.request())).rejects.toThrow("reach revoked");
expect(
await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.agentId, f.agentId)),
).toEqual([]);
});
it.each([
{ provider: "slack", state: "active", callerMarkers: false },
{ provider: "telegram", state: "completed", callerMarkers: false },
{ provider: "slack", state: "active", callerMarkers: true },
] as const)(
"rejects generic failed-chat retry before admission ($provider/$state; caller markers: $callerMarkers)",
async ({ provider, state, callerMarkers }) => {
// Keep the fixture-owned unrelated run occupying the only execution slot.
// This exercises real heartbeat admission without starting an adapter.
const f = await fixture();
const applicationId = randomUUID();
const connectionId = randomUUID();
const endpointId = randomUUID();
const conversationId = randomUUID();
const originalCommentId = randomUUID();
const failedCommentId = randomUUID();
const taskKey = `D${f.companyId.slice(0, 7)}-1`;
await db
.update(issues)
.set({
identifier: taskKey,
description: "Original request A: describe the original photo.",
})
.where(eq(issues.id, f.issueId));
await db.insert(toolApplications).values({
id: applicationId,
companyId: f.companyId,
name: "Failed chat retry",
type: "chat",
});
await db.insert(toolConnections).values({
id: connectionId,
companyId: f.companyId,
applicationId,
name: `${provider} failed chat retry`,
uid: `chat-retry-${connectionId}`,
connectionPurpose: "channel",
transport: "chat_sdk",
enabled: true,
status: "active",
});
await db.insert(chatEndpoints).values({
id: endpointId,
companyId: f.companyId,
connectionId,
provider,
publicId: randomUUID(),
assignedAgentId: f.agentId,
status: "active",
});
await db.insert(chatConversations).values({
id: conversationId,
companyId: f.companyId,
endpointId,
issueId: f.issueId,
externalConversationId: `${provider}-retry-dm`,
externalThreadId: `${provider}-retry-thread`,
externalLabel: `${provider} retry fixture`,
sessionGeneration: 1,
isDirectMessage: true,
state: "active",
});
await db.insert(issueComments).values([
{
id: originalCommentId,
companyId: f.companyId,
issueId: f.issueId,
authorUserId: "board-user",
body: "Original request A: describe the original photo.",
},
{
id: failedCommentId,
companyId: f.companyId,
issueId: f.issueId,
authorUserId: "board-user",
body: "Later request B: explain the new queue behavior, not the photo.",
},
]);
await db.insert(chatMessageLinks).values(
[originalCommentId, failedCommentId].map((commentId, index) => ({
companyId: f.companyId,
endpointId,
conversationId,
commentId,
providerMessageId: `${provider}-retry-${index}`,
direction: "inbound",
})),
);
const request = f.request(failedCommentId);
await queueIssueAssignmentWakeup({
heartbeat: f.heartbeat,
issue: {
id: f.issueId,
assigneeAgentId: f.agentId,
status: "in_progress",
},
reason: "External chat message received",
mutation: "chat_message_received",
contextSource: `chat:${provider}`,
taskKey,
requestedByActorType: request.requestedByActorType,
requestedByActorId: request.requestedByActorId,
wakeCommentId: failedCommentId,
durableChatRequest: request,
rethrowOnError: true,
});
const [admitted] = await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.wakeupRequestId, request.id));
expect(admitted).toMatchObject({
status: "queued",
contextSnapshot: {
taskKey,
source: `chat:${provider}`,
wakeCommentId: failedCommentId,
wakeCommentIds: [failedCommentId],
},
});
// Simulate a transient execution failure, not an integrity fault or a
// manual edit to any live runner root. The admitted request stays intact.
await db
.update(heartbeatRuns)
.set({ status: "failed", finishedAt: new Date() })
.where(eq(heartbeatRuns.id, admitted.id));
await db
.update(agentWakeupRequests)
.set({ status: "failed", finishedAt: new Date() })
.where(eq(agentWakeupRequests.id, request.id));
await db
.update(issues)
.set({
status: "blocked",
executionRunId: null,
executionAgentNameKey: null,
executionLockedAt: null,
})
.where(eq(issues.id, f.issueId));
if (state === "completed") {
await db
.update(chatConversations)
.set({ state: "completed" })
.where(eq(chatConversations.id, conversationId));
}
const snapshot = async () => ({
runs: await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.agentId, f.agentId))
.orderBy(heartbeatRuns.id),
receipts: await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.agentId, f.agentId))
.orderBy(agentWakeupRequests.id),
issue: await db.select().from(issues).where(eq(issues.id, f.issueId)),
conversation: await db
.select()
.from(chatConversations)
.where(eq(chatConversations.id, conversationId)),
publications: await db
.select()
.from(chatPublications)
.where(eq(chatPublications.conversationId, conversationId)),
});
const before = await snapshot();
// This is the existing IssueDetail/Inbox generic wakeup API, deliberately
// lacking a failed-run ID. It must not silently retry A in a UUID-keyed
// session after losing B's exact chat comment and admission authority.
const rejection = await f.heartbeat
.wakeup(f.agentId, {
source: "on_demand",
triggerDetail: "manual",
reason: "retry_failed_run",
payload: {
issueId: f.issueId,
...(callerMarkers
? {
retryOfRunId: admitted.id,
taskKey,
commentId: failedCommentId,
}
: {}),
},
...(callerMarkers
? {
contextSnapshot: {
source: `chat:${provider}`,
retryOfRunId: admitted.id,
wakeCommentId: failedCommentId,
wakeCommentIds: [failedCommentId],
},
}
: {}),
requestedByActorType: "user",
requestedByActorId: "board-user",
})
.then(
() => null,
(error: unknown) => error,
);
expect.soft(rejection).toMatchObject({
status: 409,
details: { code: "chat_failed_run_retry_requires_authorized_context" },
});
const after = await snapshot();
expect.soft(after.receipts).toEqual(before.receipts);
expect.soft(after.runs).toEqual(before.runs);
expect.soft(after.issue).toEqual(before.issue);
expect.soft(after.conversation).toEqual(before.conversation);
expect.soft(after.publications).toEqual(before.publications);
expect(f.authorize).toHaveBeenCalledTimes(1);
},
);
it("preserves ordinary non-chat manual retry admission", async () => {
const f = await fixture();
const retry = await f.heartbeat.wakeup(f.agentId, {
source: "on_demand",
triggerDetail: "manual",
reason: "retry_failed_run",
payload: { issueId: f.issueId },
requestedByActorType: "user",
requestedByActorId: "board-user",
});
expect(retry).toMatchObject({
status: "queued",
contextSnapshot: {
issueId: f.issueId,
taskKey: f.issueId,
wakeReason: "retry_failed_run",
},
});
const receipts = await db
.select()
.from(agentWakeupRequests)
.where(eq(agentWakeupRequests.agentId, f.agentId));
expect(receipts).toHaveLength(1);
expect(receipts[0]).toMatchObject({ status: "queued", runId: retry!.id });
expect(
await db
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.agentId, f.agentId)),
).toHaveLength(2);
});
});