feat: report the last task status when a task times out

[barnabasbusa]

When a check task runs out of time, by its own timeout or its parent's deadline, the only result was context deadline exceeded, so a failed run had to be dug out of the task logs to learn what never passed. This appends the task's last progress message to the timeout error (task timed out: context deadline exceeded, last status: Waiting for finality, finalized epoch 3 < 1000 (current epoch 5)) for every task, and makes check_consensus_finality report the condition that failed and check_consensus_sync_status name the clients that are not synced. Verified on a local devnet.

Examples

A check that never passes inside its timeout (from a local devnet run):

- name: check_consensus_finality
  timeout: 2m
  config:
    minFinalizedEpochs: 1000

before: context deadline exceeded
after:  task timed out: context deadline exceeded, last status: Waiting for finality, finalized epoch 3 < 1000 (current epoch 5) (attempt 6)

The same applies when the parent's deadline cancels the task, e.g. a check inside run_tasks_concurrent with a timeout:

before: context deadline exceeded
after:  context deadline exceeded, last status: Waiting for sync... 2/3, not synced: 3-geth-lodestar-prysm (attempt 12)

check_consensus_finality with failOnCheckMiss:

before: finality check failed
after:  finality check failed: 5 unfinalized epochs > 3 (finalized epoch 9, current epoch 14)

[redpandabot]

This change wraps task execution errors with a timeoutError helper that appends the task's last progress message when the task's own timer fired or the error wraps context.DeadlineExceeded, adds a lock-protected GetProgressMessage accessor, and makes the finality and sync-status checks report their failure conditions and unsynced clients. I traced every return path and lock interaction and found the logic, locking and error semantics consistent; the new unit test matches the helper's branches.

---

_{Reviewed 5 changed file(s) @ 7d830895 — no blocking issues found.
"RTFM."}