Skip to content

lidge-jun/codexclaw / plugins/codexclaw/skills/qa

Use for manual web, TUI, CLI, API and desktop QA. Triggers: smoke test, visual QA, screenshot check, CJK clipping, 수동 QA, 실제로 되는지 확인, 동작 확인, 직접 돌려봐.

cxc-qa — Manual Surface QA Gate

Prove a built surface works by DRIVING it, not by inferring from green tests. No scenario closes on a status string; it closes on a captured artifact from a real surface, an adversarial pass, and a teardown receipt. Everything in this skill is E7 discipline (agent-followed, not hook-enforced); the one shipped E2 touchpoint is noted in §7.

Completion proof: dev verification gate (FAMILY-PROOF-01). Every PASS points at a non-empty artifact you (or your dispatched QA worker) captured THIS round.

2. Surface detection + faithful channels

State the exact surface and invocation BEFORE running each scenario. Use the channel faithful to the surface — CLI output parsed from the wrong layer is not evidence of the layer you changed:

SurfaceFaithful channelArtifact
HTTP APIcurl -i (headers + body) — scenarios in references/http-api-qa.mdresponse capture
CLIreal invocation, stdout/stderr split + exit code — discipline in references/cli-tui-qa.mdterminal capture
TUItmux session at FIXED dims — mechanics in references/cli-tui-qa.md, visual rubric in references/visual-qa.mdplain + ANSI captures
Web UIbrowser screenshot at a STATED viewport, read via view_image — workflow in references/visual-qa.mdscreenshot(s)
Desktop GUIcomputer-use + screenshots (per-app approval; never drive terminals/Codex itself)screenshots + action log; native panels, menu-bar popups and approval prompts follow dev-devops native-desktop-acceptance.md §10 and macos-system-approvals.md

Tool choice for the browser/CU rows follows the shared portable browser policy; the inspect -> act -> re-inspect protocol applies. Data-shaped behavior may use parsed CLI/data output as its channel.

Modular References

Read the matching row before the governed action.

ConditionReference
Choosing QA ownership or closing PABCD CScope and binding
Recording any scenario verdict or aggregate receiptEvidence contract
Driving scenarios and reviewing artifactsAdversarial classes and oracle passes
Finishing any QA passTeardown receipts
ANY visual surface verdict (web UI, TUI, desktop GUI)Visual rubric
HTTP API surface scenariosWire-driving procedure
CLI or TUI surface scenariosSession mechanics
Choosing browser/CU toolsPortable browser routing (QA-TOOL-LADDER-01)
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 318kUpdated Claude CodeCodexUnverifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 318kUpdated Claude CodeCodexUnverifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 318kUpdated Claude CodeCodexUnverifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 196kUpdated Claude CodeCodexVerifiedInstall