[pascalandy]

While another call ran a transcript, its dry run under the same --run-id did not say "busy". It read the record without the lock and answered RESUME_UNSAFE (exit 1), or planned and started uv. The call itself answers TEMPORARY (exit 75), so an agent read the dry run as "never repeat" when it only had to wait.

The dry run now probes the run lock first, as the agent extension does. A busy run answers TEMPORARY with exit 75, run_id, and the hint ncly run view <id>, and starts no program. The probe creates no lock file and keeps no lock, and the dry run still writes no record. Card XT-102-T1, the last todo card of XT-102, so XT-102 is Shipped and leaves the queue.

Proof: the new steps of transcript_contenders.txtar failed before the fix (RESUME_UNSAFE, exit 1, while a real call held b1) and pass after; just check exit 0; verify-nicely by another agent passed.

• 1. Spec: extensions/transcript/spec/records-and-replay.md gives the dry run's lock probe and its TEMPORARY answer
• 2. Scenario: transcript_contenders.txtar runs the dry run while a real call holds b1. It failed before the fix with RESUME_UNSAFE, exit 1
• 3. Code: the transcript scenarios pass
• 4. Text: n/a, the answer reuses the catalog entry transcript.run.busy
• 5. Checks: just check passes
• 6. Verify: an agent that did not write the change ran verify-nicely on transcript and part of transcript-run at edb249e, PASS
• 7. Docs: XT-102-T1 is done and XT-102 is Shipped; the transcript skill teaches the busy dry run. README.md shows nothing new
• 8. Sign off: just signoff passed on 6437f8e, rebased on main after the stack ✨ feat: template: a build stops at 5 rounds, and a ledger keeps every finding #418, ✨ feat: template: finish proves the first failing run and one state of open work, and release runs finish again #421, ✨ feat: template: one case matrix before the first review, and signoff runs the first failing run again #424

Evidence of verify-nicely (run 20261101-192614-378169, paths shortened):

$ ncly transcript run youtube --url https://www.youtube.com/watch\?v=EIEc43CxIvY --output <run>/scratch/home/videos --run-id v1 --dry-run --json
{"ok":true,"contract_version":1,"dry_run":true,...,"steps":[{"id":"transcribe","action":"run"}],...}
--- exit 0
(after it: no runs/v1.lock, no runs/v1.json)

(flock <run>/scratch/home/.local/state/nicely/runs/v2.lock sleep 60 &)
$ ncly transcript run youtube --url https://www.youtube.com/watch\?v=EIEc43CxIvY --output <run>/scratch/home/videos --run-id v2 --dry-run --json
--- stderr
{"ok":false,"contract_version":1,"run_id":"v2","errors":[{"code":"TEMPORARY","message":"Another process runs v2 now.","hint":"ncly run view v2","next":["ncly","run","view","v2"]}]}
--- exit 75
(a ps poller saw no uv and no Python during this call; it saw both during the free dry runs)

(after the holder ended)
$ ncly transcript run youtube ... --run-id v2 --dry-run --json
{"ok":true,"contract_version":1,"dry_run":true,...}
--- exit 0

Reviewed with codex-loop: 1 round with gpt-6.1-sol. Sol verdict: approved

Created with Claude Opus 5.5 in Claude Code.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: Repository UI
• Review profile: ASSERTIVE
• Plan: Advanced
• Run ID: db6f0806-f374-445c-bd9b-de413739d28b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• Trigger review

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts