Agent skill · software engineering · thienanblog
testing-verification
Coordinator-routed specialist when test strategy, coverage, QA, acceptance verification, CI-check design, browser verification, Playwright E2E, or visual comparison is the primary work. Use after project-development-mindset routes here, or directly when explicitly invoked or installed standalone. Do not use merely because implementation needs focused tests; unexplained failures route to debugging first.
Why this skill is useful
Adds specific test strategies and commands for Playwright and CI that the AI wouldn't reliably generate on its own.
What it needs
Requires playwright installed locally. About 2k tokens when loaded. Last updated 2026-08-05. 63 stars on the source repository.
What this skill does
Testing Verification Prefer project evidence over generic test advice, and verify behavior at the narrowest reliable level before broadening. Run this skill in the main conversation. Do not spawn subagents, agent teams, or delegated parallel workers unless the user explicitly approves the proposed count and scope after being told that doing so can increase usage. Ask again before expanding an approved scope. Operating Rules Follow project instructions, existing test tools, naming conventions, and nearby tests. Test public behavior and contracts at the narrowest reliable level; avoid implementation details unless the project expects them. Do not add a new testing framework unless the project has no reasonable existing path and the benefit is clear. Keep tests maintainable: small setup, clear assertions, stable fixtures, no hidden network or production dependencies. Workflow 1. Discover And Scope Identify test commands, CI jobs, nearby tests, browser tooling, fixtures, helpers, and source-of-truth behavior docs. Read references/test-strategy.md, choose the lowest reliable level, and prioritize targeted checks before broader suites. Use documented manual steps only when automation is unavailable or disproportionate. 2. Add Tests Or Checks Reuse existing helpers and fixtures; keep each test focused on one coherent behavior or scenario. Include edge cases that are business-relevant, not exhaustive noise. For bug fixes, add a regression test that fails before the fix when practical. For docs-only changes, verify links, examples, generated outputs, or commands when relevant. 3. Verify UI In A Browser For interactive, exploratory, screenshot, or visual-comparison work, use the harness's built-in Browser when available and follow its instructions. Use Playwright MCP only when built-in Browser is unavailable, and record why. If the user explicitly selected Browser, report the blocker and ask before substituting another surface. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills thienanblog/testing-verification--606ab2