Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →butterbase-ai/journey-deploy
Use as the deploy-verification stage of the Butterbase journey, after journey-frontend (or after any build stage if there is no frontend). Smoke-tests the deployed app — hits the live URL, invokes any deployed functions, checks auth round-trip. Writes results to docs/butterbase/04-build-log.md. Blocks journey-submit if smoke fails.
Passed 10 of 17 test cases with the skill, and 3 of 17 without it.
and 9 more in the full report
10 fixed3 passed either way8 still failing1 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 9 test cases the skill fixed. Same model, same prompt, with and without the skill.
Our docs/butterbase/03-preflight.md file was generated 30 hours ago, and docs/butterbase/00-state.md has app_id: "app_12345". We want to run deploy verification. What is the immediate first action?
| Checked | Without skill | With skill |
|---|---|---|
| Invoke butterbase-skills:journey-preflight when 03-preflight.md is older than 24 hours. | Missed | Passed |
The instructions your model reads when the skill loads.
Stage 4 of the guided journey. Verify the deployed app actually works end-to-end.
journey when current_stage: deploy./butterbase-skills:journey-deploy.If docs/butterbase/03-preflight.md is missing, older than 24 hours, or 00-state.md has app_id: null, invoke butterbase-skills:journey-preflight first. Wait for it to return successfully before proceeding.
docs/butterbase/00-state.md — for app_id, deployed_url.docs/butterbase/02-plan.md — for the function list to smoke.docs/butterbase/04-build-log.md — to know what shipped.butterbase_docs with topic: "frontend". For end-to-end smoke patterns, also WebFetch https://docs.butterbase.ai/deploy. Skip if cache is fresh.For each check, log a one-line result to 04-build-log.md. Halt on the first ✗ and ask the user to fix or skip.
Frontend reachable. If deployed_url is set, fetch it via Bash curl -sS -o /dev/null -w "%{http_code}" <url>. Expect 200. Log <ISO> deploy curl 200 ok or ... <code> fail.
Functions respond. For each function in the plan with HTTP trigger, call mcp__butterbase__invoke_function with a representative payload. Expect 2xx. Log per function.
Auth round-trip (if configured). If the plan included OAuth, prompt the user: "Open <deployed_url>, click 'Sign in with <provider>', and confirm you land logged-in. Did it work? (yes/no)". Log the answer.
Function logs clean. Call mcp__butterbase__manage_function action: get_logs for the most recent invocations. Show the user any error-level lines. Ask: "Anything concerning here? (no → continue, yes → fix and re-run)".
If all checks pass, tick - [x] deploy in 00-state.md, set current_stage: submit (if hackathon_mode: true) or current_stage: done (otherwise).
Return to journey orchestrator.
04-build-log.md.test cases passed, vs 3 / 17 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.