smoke-test

Smoke Test

You are running a post-sprint validation. Your job is to verify the application still works correctly after recent changes.

Step 1 — Understand what changed

Read the latest sprint plan to understand what was modified: tasks/sprints/sprint-*/plan.md (latest).

If tasks/sprints/sprint-*/build-output.md exists, read it for the list of modified files.

Step 2 — Run the test suite

Detect the test runner:

• If package.json exists → npm test or check scripts
• If pytest.ini, setup.cfg, pyproject.toml → pytest
• If Makefile with test target → make test
• If Cargo.toml → cargo test
• If none found → report "no test runner detected"

Run the full test suite and capture results.

Step 3 — Compare with baseline

Check tasks/lessons.md or previous sprint outputs for baseline test counts. If a baseline exists, compare:

• Are there fewer tests now? (tests removed?)
• Are there new failures?

Step 4 — Health checks

If the project has identifiable health endpoints or CLI commands:

• Web app: check if key endpoints respond (look for routes in code)
• CLI: verify key commands run without error
• API: check health/status endpoint if it exists

If none are identifiable, skip this step and note it.

Step 5 — Check for common regressions

• Imports that break (look for import errors in test output)
• Config files that changed (diff against previous sprint)
• Dependency conflicts (check lock file changes)

Step 6 — Produce output

Save to tasks/sprints/sprint-XX/smoke-test-output.md:

# Smoke Test — Sprint [XX]

## Test suite
- Runner: [pytest/npm/etc.]
- Total tests: [count]
- Passed: [count]
- Failed: [count]
- Baseline (last sprint): [count if known]

## Regression check
- New test failures: [list or "none"]
- Missing tests (removed?): [list or "none"]

## Health checks
- [Endpoint/command]: [status]

## Verdict
[PASS / FAIL / PASS WITH WARNINGS]
[If fail: what specifically failed and recommended action]

Channel mode (remote execution)