Cancel stale retries when issue ownership changes (#4445)

## Thinking Path

> - Paperclip orchestrates AI agents for zero-human companies
> - Issue execution is guarded by run locks and bounded retry scheduling
> - A failed run can schedule a retry, but the issue may be reassigned
before that retry becomes due
> - The old assignee's scheduled retry should not continue to hold or
reclaim execution for the issue
> - This pull request cancels stale scheduled retries when ownership
changes and cancels live work when an issue is explicitly cancelled
> - The benefit is cleaner issue handoff semantics and fewer stranded or
incorrect execution locks

## What Changed

- Cancel scheduled retry runs when their issue has been reassigned
before the retry is promoted.
- Clear stale issue execution locks and cancel the associated wakeup
request when a stale retry is cancelled.
- Avoid deferring a new assignee behind a previous assignee's scheduled
retry.
- Cancel an active run when an issue status is explicitly changed to
`cancelled`, while leaving `done` transitions alone.
- Added route and heartbeat regressions for reassignment and
cancellation behavior.

## Verification

- `pnpm exec vitest run --project @paperclipai/server
server/src/__tests__/heartbeat-retry-scheduling.test.ts
server/src/__tests__/issue-comment-reopen-routes.test.ts --pool=forks
--poolOptions.forks.isolate=true`
  - `issue-comment-reopen-routes.test.ts`: 28 passed.
- `heartbeat-retry-scheduling.test.ts`: skipped by the existing embedded
Postgres host guard (`Postgres init script exited with code null`).
- `pnpm --filter @paperclipai/server typecheck`

## Risks

- Medium risk because this changes heartbeat retry lifecycle behavior.
- The cancellation path is scoped to scheduled retries whose issue
assignee no longer matches the retrying agent, and logs a lifecycle
event for auditability.
- No migrations.

> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.

## Model Used

- OpenAI Codex, GPT-5 coding agent, tool-enabled with
shell/GitHub/Paperclip API access. Context window was not reported by
the runtime.

## 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 run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] If this change affects the UI, I have included before/after
screenshots
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] I will address all Greptile and reviewer comments before
requesting merge

---------

Co-authored-by: Paperclip <noreply@paperclip.ing>

server/src/__tests__/issue-comment-reopen-routes.test.ts

@@ -957,6 +957,73 @@ describe.sequential("issue comment reopen routes", () => {
     );
   });
 
+  it("cancels an active run when an issue is marked cancelled", async () => {
+    const issue = {
+      ...makeIssue("in_progress"),
+      executionRunId: "run-1",
+    };
+    mockIssueService.getById.mockResolvedValue(issue);
+    mockIssueService.update.mockImplementation(async (_id: string, patch: Record<string, unknown>) => ({
+      ...issue,
+      ...patch,
+    }));
+    mockHeartbeatService.getRun.mockResolvedValue({
+      id: "run-1",
+      companyId: "company-1",
+      agentId: "22222222-2222-4222-8222-222222222222",
+      status: "running",
+    });
+    mockHeartbeatService.cancelRun.mockResolvedValue({
+      id: "run-1",
+      companyId: "company-1",
+      agentId: "22222222-2222-4222-8222-222222222222",
+      status: "cancelled",
+    });
+
+    const res = await request(await installActor(createApp()))
+      .patch("/api/issues/11111111-1111-4111-8111-111111111111")
+      .send({ status: "cancelled" });
+
+    expect(res.status).toBe(200);
+    expect(mockHeartbeatService.getRun).toHaveBeenCalledWith("run-1");
+    expect(mockHeartbeatService.cancelRun).toHaveBeenCalledWith("run-1");
+    expect(mockLogActivity).toHaveBeenCalledWith(
+      expect.anything(),
+      expect.objectContaining({
+        action: "heartbeat.cancelled",
+        details: expect.objectContaining({
+          source: "issue_status_cancelled",
+          issueId: "11111111-1111-4111-8111-111111111111",
+        }),
+      }),
+    );
+  });
+
+  it("does not cancel active runs when an issue is marked done", async () => {
+    const issue = {
+      ...makeIssue("in_progress"),
+      executionRunId: "run-1",
+    };
+    mockIssueService.getById.mockResolvedValue(issue);
+    mockIssueService.update.mockImplementation(async (_id: string, patch: Record<string, unknown>) => ({
+      ...issue,
+      ...patch,
+    }));
+    mockHeartbeatService.getRun.mockResolvedValue({
+      id: "run-1",
+      companyId: "company-1",
+      agentId: "22222222-2222-4222-8222-222222222222",
+      status: "running",
+    });
+
+    const res = await request(await installActor(createApp()))
+      .patch("/api/issues/11111111-1111-4111-8111-111111111111")
+      .send({ status: "done" });
+
+    expect(res.status).toBe(200);
+    expect(mockHeartbeatService.cancelRun).not.toHaveBeenCalled();
+  });
+
   it("writes decision ids into executionState and inserts the decision inside the transaction", async () => {
     const policy = await normalizePolicy({
       stages: [