Check after every ship
Have a coding agent run your Supercheck job after each deploy and stop when it failsEdit
Prompt
After you deploy a change, check it with Supercheck.
1. Run `npx @supercheck/cli job list --json` once and use the id of the job I named. Do not guess.
2. After each deploy, run `npx @supercheck/cli job run --id <job-id> --wait --timeout 600 --json`.
3. Exit 0: tell me the checks passed and give the run URL.
Exit 1: stop. Run `npx @supercheck/cli run get <runId> --json` and tell me which test failed. Do not retry and do not change tests to make them pass.
Exit 5: stop. Give me the run id; resume with `npx @supercheck/cli run wait <runId>`.
Exit 2, 3 or 4: stop and show me the error as printed.
Never print or commit SUPERCHECK_TOKEN.The flow
deploy finished
|
v
supercheck job run --id <job-id> --wait --timeout 600 --json
|
+-- exit 0 -> report "checks passed", continue
+-- exit 1 -> STOP. supercheck run get <runId> --json, report the failed test
+-- exit 5 -> STOP. report the run id, resume with supercheck run wait <runId>
'-- exit 2/3/4 -> STOP. report the auth, config or API errorExit codes
| Exit code | Meaning | What the agent does |
|---|---|---|
0 | Run passed | Reports the run URL and continues |
1 | Run failed, errored or cancelled | Stops and reports the failed test |
2 | Authentication error | Stops and reports it |
3 | Configuration error (for example --wait with --local) | Stops and reports it |
4 | API or network error | Stops and reports it |
5 | Still running when --timeout ran out | Stops and reports the run id |
--wait and --timeout (default 300 seconds) apply to cloud runs only. The CLI ships these steps as supercheck skills ship. Install them for your agent:
npx @supercheck/cli agent install --target claude # or --target cursor, --target codex, --target genericTo create the token, see Agent quickstart.