After every production deploy, fire a Claude Code task that runs smoke checks, scans error logs, and posts a go / no-go to the release channel before the window closes: AI agent prompt
Anthropic’s Claude Code scheduled-tasks documentation lists deploy verification as a first-class example: the CD pipeline calls the task’s API /fire endpoint after each production deploy; the task runs smoke checks against the new build, scans error logs for regressions, and posts a go or no-go to the release channel before the deploy window closes. Pair it with branch protection and a human owner for rollback decisions. Feature is in research preview—confirm availability and usage limits on your plan before relying on it in a release train.
The prompt
1. Write a one-page smoke checklist (health URL, login, one write path, error-rate threshold) and the exact go / no-go message format for your release channel. 2. Create a Claude Code cloud task with an API trigger whose prompt runs only that checklist, reads recent error logs, and posts the verdict—never rolls back by itself. 3. From your CD pipeline, POST to the Fire URL right after the production deploy succeeds, including the release id and URLs in the text payload. 4. Dry-run on staging until go/no-go matches what a human would have said. 5. Keep a named human as release owner: they read the verdict and trigger rollback if needed; the agent only reports.
How to use it
- Copy the prompt above.
- Replace every [bracket] with your own details (or fill in the blanks on AgentPlaybook Real Estate, which remembers your answers for the next prompt).
- Paste it into Claude Code cloud scheduled tasks (API trigger), CD/CI pipeline, app health endpoints + log store and review the answer before you use it.
Works with: Claude Code cloud scheduled tasks (API trigger), CD/CI pipeline, app health endpoints + log store, Slack/Teams release channel
Real-world example: code.claude.com/docs/en/routines
Before you use it: what to check
Already built into this prompt
- Create a Claude Code cloud task with an API trigger whose prompt runs only that checklist, reads recent error logs, and posts the verdict—never rolls back by itself.
- Dry-run on staging until go/no-go matches what a human would have said.
- Keep a named human as release owner: they read the verdict and trigger rollback if needed; the agent only reports.
Check yourself
- Read and edit the draft before it reaches a customer or goes public. If you later automate this prompt, start with "draft and notify me", not "send automatically".
- Check names, dates, numbers and any claim against your own source material before you use the answer. AI can misread or invent details.
Using an AI agent? Download this prompt as a SKILL.md agent skill with these checks built in, or get the .zip for Claude.ai. How to install skills.
Want to make this prompt even better? Grade it with the free prompt grader or build your own with the AI prompt generator.
32 copy-paste agent prompts (PDF) plus 5 new prompts every week.
Get it free