lidge-jun/codexclaw / plugins/codexclaw/skills/qa
cxc-qa
VerifiedUse 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.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 3.1 KB |
| agents/openai.yaml | config | 155 B |
| references/adversarial-oracles.md | markdown | 2.5 KB |
| references/cli-tui-qa.md | markdown | 3.2 KB |
| references/evidence-contract.md | markdown | 3.7 KB |
| references/http-api-qa.md | markdown | 2.9 KB |
| references/scope-and-binding.md | markdown | 1.3 KB |
| references/teardown-receipts.md | markdown | 392 B |
| references/visual-qa.md | markdown | 7.1 KB |
| scripts/validate-evidence.mjs | script | 21.1 KB |
references/cli-tui-qa.md
View on GitHub# CLI / TUI QA — session driving + capture mechanics (QA-CLI-01)
How to DRIVE terminal surfaces and capture honest evidence. E7 discipline.
Role split: this file owns session/capture MECHANICS; the visual artifact
rubric and harness options (VHS/teatest, width gates as visual findings) stay
in `references/visual-qa.md`.
> These examples use POSIX conventions. On Windows: use `NUL` instead of
> `/dev/null`, `taskkill` instead of `kill`, and PowerShell equivalents for
> `awk`/`tmux` workflows.
## CLI capture discipline
- Capture stdout and stderr SEPARATELY when the claim distinguishes them
(`cmd >out.txt 2>err.txt`); `2>&1` merges and destroys that evidence. Also
capture the merged stream when ordering matters. ALWAYS record the exit
code (`echo "exit=$?"`) in the same artifact.
- **Smoke floor** — `--help` and `--version` exit 0 and print to stdout (help
requested is not an error); an UNKNOWN flag exits nonzero and names the
flag on stderr.
- **TTY vs pipe** — run the claim's mode for real: `cmd | cat` forces
non-TTY; interactive/TTY behavior runs inside tmux (below). Color/spinner
output that corrupts piped consumption is a finding.
- **Env/locale matrix** (when behavior claims depend on it) — `NO_COLOR=1`,
`TERM=dumb`, and a non-default `LANG`/`LC_ALL`; capture each run
separately.
- **Config precedence** — when the tool reads flags + env + config file,
drive one conflicting combination and capture which source won; silent
mystery-precedence is a finding.
- **stdin behavior** — empty stdin (`</dev/null`) must not hang; piped input
path captured when the tool claims it.
- **Signal/cleanup** — for long-running commands, SIGINT mid-run, then
capture the cleanup receipt (no orphan processes via `ps`, no stale
lockfiles/temp dirs). Pair with SKILL.md §6 teardown.
## TUI session mechanics (tmux)
Deterministic lifecycle — never drive a TUI in your own terminal:
1. `tmux new-session -d -s qa-<id> -x <cols> -y <rows>` — FIXED dims, stated
in the matrix (defaults: 80x24 + one narrow run, e.g. 40 cols, as the §4
narrow class).
2. `tmux send-keys -t qa-<id> '<cmd>' Enter` — one action per send.
3. **Wait for a marker, never sleep-and-hope**: poll
`tmux capture-pane -pt qa-<id>` until an expected string appears (bounded
retries, then FAIL with the last capture as the blocker artifact).
4. Capture evidence: `capture-pane -p` (plain) and `-e` (ANSI) copies, taken
at each asserted state, numbered per step.
5. `tmux kill-session -t qa-<id>` + `tmux ls` clean output = the teardown
receipt (SKILL.md §6).
Interaction claims (keystrokes, resize) re-inspect after every action —
inspect -> act -> re-inspect, same protocol as browser QA. Resize behavior:
`tmux resize-window -t qa-<id> -x <n>` then re-capture; a redraw that leaves
stale fragments is a finding.
Width/overflow and wide-char (CJK) column checks run on the plain captures
(`awk 'length > COLS'` minimum gate); their RUBRIC and harness escalation
(VHS/teatest) live in `references/visual-qa.md` §TUI.
## Artifacts
Per SKILL.md §3: `invocation.txt` carries the full tmux/shell command
sequence (copy-pasteable); captures are numbered per step; `verdict.json`
per scenario. Env-matrix runs are separate scenarios, not overwrites.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| 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: 197k | Updated | Claude CodeCodex | Verified | Install |