菜单
Use when designing test suites, writing acceptance tests, validating an implementation against a spec, or planning test coverage before implementation. Trigger: /qa-engineer.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
Use when a diff or PR is too big to review by eye and the question is "what should I actually look at" — sorts each changed line into wire / CLI / API / logic / test / doc tiers, then opens a clickable local page with the real hunks. A reading order, never a verdict.
Use to implement a plan artifact from `/swarm-plan`, or a free-text implementation task with contract-first TDD + Review-Fix Loop. Tier (`low | auto | high | max`) scales builder model, loop rounds, review breadth, and Codex code-diff gate.
Use for branch, PR, or diff review before landing on main. Tier (`low | auto | high | max`) scales breadth from single reviewer to adversarial panel with architect + SOTA-gap + Codex cross-model gate. Triggers: "review", "/swarm-review", pre-merge checks.
Use when the user says "commit", "/commit", "save progress", or asks to land working changes on a feature or worktree branch. OCX working-phase posture — minimises commit count, can amend rolling Checkpoints, offers a one-time PR prompt. For pre-merge cleanup, use `/finalize`.
Use for feature planning, task decomposition, multi-perspective research, or ADR scaffolding. Tier (`low | auto | high | max`) scales research depth, architect model, and review breadth. Canonical research primitive for AI config / ADR work.
Use when the user says "codex review", "/codex-adversary", "adversarial review", "cross-model review", or asks for a different-model second opinion on a diff or plan artifact. Complements `/swarm-review` via Codex-CLI. Scopes: code-diff (default), plan-artifact (gate for `/swarm-plan`).
| name | qa-engineer |
| description | Use when designing test suites, writing acceptance tests, validating an implementation against a spec, or planning test coverage before implementation. Trigger: /qa-engineer. |
Role: test strategy, writing, validation for OCX.
Tests written before implementation from design record:
unimplemented!() / NotImplementedError against stubsAnalyze → plan → write → run → cover happy, error, edge cases.
.claude/rules/subsystem-tests.md — pytest fixtures (ocx, published_package, unique_repo), OcxRunner API, make_package(), assertion helpers (auto-loads on test/**).claude/rules/quality-core.md + quality-python.md — test quality, async patterns.claude/rules/quality-rust.md — unit test patterns for ocx_libtask runner — task test:quick, task test:parallel, task test:unit, task coverage. Never ad-hoc pytest/cargo when task exists.assert_symlink_exists() not path.is_symlink() (Windows junction compat)$ARGUMENTS