Files
2025-11-30 08:37:27 +08:00

1.7 KiB

name, description
name description
ln-343-manual-tester Performs manual testing of Story AC (API/UI) with curl/puppeteer, documents results in Linear (structured comment) and provides reusable scripts. Worker only.

Manual Tester

Manually verifies Story AC on running code and reports structured results for the quality gate.

Purpose & Scope

  • Rebuild/run app, detect API/UI, and execute AC-driven checks via curl/puppeteer.
  • Document results in Linear (Format v1.0) with pass/fail per AC, edge/error handling, and temp script path.
  • No status changes or task creation.

Workflow (concise)

  1. Env setup: Rebuild containers (no cache), start, ensure healthy. Detect API vs UI; confirm app reachable.
  2. Load AC: Fetch Story, parse AC into Given/When/Then list (3-5 expected).
  3. Execute: For each AC + edge/error cases, run curl (API) or puppeteer (UI); capture responses/screens; note deviations.
  4. Report: Produce verdict PASS/FAIL; add Linear comment (Format v1.0) with AC matrix, issues found, evidence, and temp script path; return JSON result.

Critical Rules

  • Rebuild Docker before testing; fail if rebuild/run unhealthy.
  • Keep language of Story (EN/RU) in comment.
  • No fixes or status changes; only evidence and verdict.

Definition of Done

  • App rebuilt and running; AC parsed.
  • Tests executed across main + edge/error cases.
  • Verdict and structured Linear comment posted with evidence and script path.

Reference Files

  • AC format and scripts: ../ln-350-story-test-planner/references/test_task_template.md (for alignment)
  • Risk-based context: ../ln-350-story-test-planner/references/risk_based_testing_guide.md

Version: 3.0.0 (Condensed manual testing flow) Last Updated: 2025-11-26