[factory] fleet.test.ts "keeps relay dispatch ownership…" times out at 5s on a loaded runner, failing CI on main

[khaliqgant]

The failure

FAIL src/cli/fleet.test.ts > "keeps relay dispatch ownership until the remote PR is published and the issue is parked"
Test timed out in 5000ms

It is not any one PR's regression — it fails on main

Measured by slot-wedge-r2-0902 while trying to land #440, and worth recording
because "CI is red" has been read as a blocker on individual PRs several times
today:

• main @ 3d7561a, with none of that lane's changes: the same test fails,
  same 5000ms timeout.
• The lane's branch, run 1: the same test, same timeout.
• The lane's branch, run 2: the same test plus a different one (a Slack
  reply-route test).

Each rerun failing a different test is the signature of a load flake, not a
regression. The lane's diff touched only src/orchestrator/factory.ts and
factory.test.ts; fleet.test.ts is a file it does not touch.

The mechanism

The suite has no testTimeout configured, so it runs at Vitest's 5s
default. This test sits in a 286-second suite on a machine also running a
fleet of agents, and a loaded runner trips the default.

Why it matters

It makes the package check red on unrelated PRs, so a genuinely-failing PR and
a flake-failing PR look identical. That has already cost review time tonight,
and it trains readers to wave CI through — which is how a real failure gets
merged.

Suggested fix

Give this test (or the suite) a testTimeout proportionate to what it actually
does, rather than the 5s default. If 5s is genuinely the intended budget, the
test needs to stop depending on wall-clock scheduling under load.

Please do not fix it by deleting or skipping the assertion — it covers
dispatch ownership through PR publication, which is exactly the area under
active repair (#430, #440).