Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add tamnguyendinh/Anvien --skill qagit clone --depth 1 https://github.com/tamnguyendinh/AnvienWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/tamnguyendinh/anvien/qa)<a href="https://agentmods.dev/skills/tamnguyendinh/anvien/qa"><img src="https://agentmods.dev/badge/skills/tamnguyendinh/anvien/qa/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/tamnguyendinh/anvien/qa"><img src="https://agentmods.dev/badge/skills/tamnguyendinh/anvien/qa.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00055 | $0.01288 |
| Opus 5.5 | $0.00022 | $0.00515 |
| Sonnet 5.5 | $0.00011 | $0.00258 |
| Haiku 4.5 | $0.00006 | $0.00129 |
Grade A, and why
qa scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- Backend/API tools: use HTTP clients (curl/fetch), database CLI tools, or test runners for non-UI scopes. How it starts
The opening of the file, as written. The whole thing — 91 lines — stays where its author put it; the contents beside it link to each section on GitHub.
(MUST) For Codex: When QA must use plugins
- Browser - Control the in-app browser
- Chrome - Control the user's real Chrome browser
- Computer Use - Control Windows apps or installed artifacts when QA requires real app interaction outside a browser.
- Playwright: use as an automation arm for browser actions, control sweeps, screenshots, videos, traces, and reports.
- Backend/API tools: use HTTP clients (curl/fetch), database CLI tools, or test runners for non-UI scopes.
(MUST) For Claude or other agents:
- Use the equivalent browser, Chrome/session, computer-control, or HTTP/database CLI capability exposed by that agent/runtime.
QA Runtime Review
Use this skill to verify real runtime behavior from the user's point of view. QA is a no-fix role: inspect, execute, record, classify, and report. Do not repair product code, tests, specs, fixtures, or generated files during QA.
Iron Laws
- QA MUST NOT fix code. Report bugs, blockers, friction, test gaps, and suggested fix direction only.
- Runtime behavior is the source of QA truth. Code presence, tests, screenshots, or graph output do not prove pass/fail by themselves.
- Must have screenshot Evidence for UI scopes. For non-UI / Backend / Database scopes, follow
references/backend-qa-protocol.md(structured payload, DB state diff, and execution logs). - Tests verify already-correct behavior. Tests do not define expected behavior and must not legitimize wrong UX.
- Wrong tests are findings. A green test that encodes wrong behavior is not QA proof.
- Do not skip steps. Every in-scope flow step matters equally.
- Every in-scope surface, action, state, data path, and blocker needs an explicit verdict or out-of-scope mark.
- An incomplete inventory or action ledger means QA is incomplete.
- If a blocker prevents downstream verification, downstream scope is
Blocked, not passed. - For final visible QA, open the real built website, runtime, or generated app artifact the same way a user would open it, in a visible browser or app window on the user's physical PC. Headless, hidden, source-only, test-only, or screenshot-only execution cannot approve final visible QA.
- QA must run on the real, officially built application (actual Electron artifact or release binary).
- For UI/visible scopes: Every action (clicking buttons, entering data, changing state) must capture clear PNG screenshots saved sequentially into the
screenshots/directory. - For UI/visible scopes: Visually inspect each captured image / visual artifact to detect UI and behavioral defects; never rely solely on in-code assertions or generated
.json/.mdsummary files. - Never accept a UI QA report lacking the actual screenshot evidence set for each interaction. For Backend / non-UI reports, require structured HTTP status, response payload, and DB state diff evidence per
references/backend-qa-protocol.md.
What ships with it
9 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- references/backend-qa-protocol.md 4.2 KB
- references/e2e-data-flow.md 1.2 KB
- references/evidence-and-blockers.md 1.8 KB
- references/frontend-automation-and-control-sweep.md 2.7 KB
- references/frontend-coverage-model.md 2.6 KB
- references/frontend-i18n-and-locale.md 490 B
- references/frontend-runtime-and-build-rules.md 5.8 KB
- references/reporting.md 4.4 KB
- references/source-of-truth-rules.md 1.3 KB
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 91 lines · 55 tokens per session scan A cc8b5914357d
qa is a skill published in the GitHub repository tamnguyendinh/Anvien (12 stars, last pushed 5d ago), licensed MIT. It adds 55 tokens to every session and 1,288 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
source-command-swarm-testing
Migrated source command swarm-testing.
source-command-sparc-tester
Migrated source command sparc-tester.
stage-9-deployment
You are a QA gate. All tests must pass. No exceptions. You do not interpret failures — you report them and block. The full test suite runs: unit tests, integration tests, and any end-to-end tests defined in prd-tests.md. A single failure blocks progression.
test-master
Generates test files, creates mocking strategies, analyzes code coverage, designs test architectures, and produces test plans and defect reports across functional, performance, and security testing disciplines. Use when writing unit tests, integration tests, or E2E tests; creating test strategies or automation…
cua-driver
Use Cua Driver for desktop or browser tasks that are awkward or unavailable through Bash/APIs, or when the user explicitly wants GUI interaction: app testing, visual bug reproduction, form filling, calendar entry, screenshots, and demo recording. Also covers Cua setup; not OpenAI Codex Computer Use or web research.
test-warp-ui
Guides testing Warp UI features and changes using the computer use tool. Use this skill only when computer-use testing was requested (explicit request or accepted offer) and the computeruse tool is available to the agent. Covers launching Warp and verifying UI behavior.