AgentStack
SKILL verified MIT Self-run

Smoke Test

skill-gunesbizim-agent-smith-smoke-test · by gunesbizim

Post-PR verification — launch the app and check health, then walk the migrate/smoke/rollback checklist. Use after a merge or before a release to confirm the running app is healthy. Guidance executed by Claude Code, not by agent-smith infra.

No reviews yet
0 installs
16 views
0.0% view→install

Install

$ agentstack add skill-gunesbizim-agent-smith-smoke-test

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

Are you the author of Smoke Test? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

You verify a change actually works in the running app and guide post-merge steps. agent-smith generates this skill; Claude Code is the runtime — there is no agent-smith deploy engine.

Step 0 — Plan (smith-mode for multi-step verification)

For anything beyond a single check, follow .claude/skills/smith-mode/SKILL.md: stage map → delegate → failable verification → self-critique.

Step 1 — Launch the app

Use the project's real run command ({{FRONTEND_DEV_SERVER_CMD}} for web, or the backend's documented start command). For a web app, drive it with the Playwright / chrome-devtools MCP; for a CLI/service, invoke it and inspect output/exit code.

Step 2 — Health checks (failable)

  • Hit the health/readiness endpoint (or run the smoke command) and assert a 2xx / expected output.
  • Exercise one critical happy-path flow end-to-end.
  • A check that cannot fail is not a check — assert a concrete observable, never "looks fine".

For a web app, run these chrome-devtools checks immediately after the app loads:

mcp__chrome-devtools__list_console_messages()   # assert zero errors / uncaught exceptions
mcp__chrome-devtools__list_network_requests()   # assert no failed (4xx/5xx) API requests

Both must be clean before the health check passes. Any console error or failed network request is a no-go finding — log it in the output with the exact message and request URL.

Step 3 — Post-merge checklist (guidance)

Walk these as guidance (you execute via Claude Code; agent-smith does not own infra):

  1. Migrations — run pending DB migrations with the project's real migrate command; confirm success.
  2. Smoke — repeat Step 2 against the deployed/staging target.
  3. Rollback readiness — confirm the previous release is recoverable (tag/image/commit), and

note the exact rollback command before promoting.

Output

Report: launched (y/n), health result per check (with the observed value), checklist status, and a clear go / no-go verdict with the reason.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.