deploy-check

Deploy Check

Run a full pre-deployment verification pipeline: lint, test, build. Fix any failures automatically.

What to do

1. Run the full pipeline (lint + test + build).
2. If anything fails, read the error output, fix the issue, and re-run.
3. Repeat until all three pass.

Flow

flow:
  retry max 5
    run: npm run lint && npm test && npm run build
    if command_failed
      prompt: The pipeline failed. Read the error output, fix the issue, and try again.
    end
  end

done when:
  tests_pass
  lint_pass

Rules

• Fix the first error you encounter — later stages may depend on earlier ones.
• If a fix in a later stage breaks an earlier stage, the next retry will catch it.
• Report the final status of all three stages when done.

Environment-Specific Command Mapping

When the repo uses a non-Node stack, substitute equivalent commands:

• Python: ruff check . && pytest && python -m build
• Go: go test ./... && go vet ./...
• Rust: cargo test && cargo clippy -- -D warnings && cargo build

Use project-native commands if scripts are already defined in package/tooling config.

Deployment Readiness Checklist

Before declaring ready:

1. Lint is clean.
2. Tests are green.
3. Build/package step completes.
4. No unresolved TODO-level warnings introduced in touched files.
5. Any migration or rollout prerequisites are called out explicitly.

Failure Classification

• lint failures: style/static correctness; fix first.
• test failures: behavior regressions; isolate smallest failing unit first.
• build failures: integration/packaging/runtime typing.

If failures are flaky or environment-specific, report the exact command and stderr fragment.

Done Response Template

• Lint: pass/fail
• Tests: pass/fail
• Build: pass/fail
• Ready: yes/no
• Notes: any remaining risk or prerequisite