Use when about to claim work is complete, fixed, or passing — requires running verification commands and confirming output before making any success claims. Evidence before claims, always.
中文简介正在准备中。当前显示原文说明。
查看安装方法安装前可以查看交给 AI 代理的指令内容。
NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE
If you haven't run the verification command in this response, you cannot claim it passes. Period.
Before ANY of these:
1. IDENTIFY → What command proves this claim?
2. RUN → Execute it. Fresh. Complete. Not a cached result.
3. READ → Full output. Check exit code. Count failures.
4. VERIFY → Does the output support the claim?
5. CLAIM → Only now may you state the result.
Skip any step = the claim is unverified.
| Claim | Required Evidence | Not Sufficient |
|---|---|---|
| "Tests pass" | Full test output showing 0 failures | Previous run, "should pass", partial suite |
| "Build succeeds" | Build command output, exit code 0 | Linter passing (linter ≠ compiler) |
| "Lint clean" | Linter output showing 0 errors/warnings | Build passing (build ≠ lint) |
| "Bug fixed" | Reproduce original symptom → now passes | "I changed the code" |
| "Regression test works" | Red-green cycle: fail → fix → pass | Test passes once without verifying it can fail |
| "Requirements met" | Line-by-line checklist against acceptance criteria | "Tests pass" (tests may not cover requirements) |
| "Migration works" | Run migration up AND down successfully | "Schema looks correct" |
When claiming a regression test is valid:
1. Write test
2. Run test → MUST FAIL (Red)
3. Implement fix
4. Run test → MUST PASS (Green)
5. Revert fix temporarily
6. Run test → MUST FAIL AGAIN (proves test is meaningful)
7. Restore fix
8. Run test → Passes (final confirmation)
If the test never fails, it proves nothing.
Catch yourself if you're about to say:
| Phrase | What to Do Instead |
|---|---|
| "Should work now" | Run the command |
| "I'm confident this is correct" | Confidence ≠ evidence |
| "Looks like it should pass" | Looks ≠ verified |
| "Just this once I'll skip" | No exceptions |
| "The linter passed so the build should too" | They check different things |
| "I changed the right file" | Run the test |
| "The agent said it succeeded" | Verify independently |
This skill reinforces the pipeline's completion-checklist rule:
Run the command. Read the output. Count the failures. THEN make the claim.
This is non-negotiable. The cost of verification is 30 seconds. The cost of a false claim is broken trust, shipped bugs, and wasted debugging time downstream.
还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问，或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
Audit and improve web accessibility following WCAG 2.1 guidelines. Use when asked to "improve accessibility", "a11y audit", "WCAG compliance", "screen reader support", "keyboard navigation", or "make accessible".
中文简介正在准备中。当前显示原文说明。
Structured methodology for adversarial thinking — generating attack scenarios, abuse cases, race conditions, and security edge cases against specs and implementations. Produces spec-level gap items, not code-level fixes.
中文简介正在准备中。当前显示原文说明。
Orchestrate multiple Antigravity skills through guided workflows for SaaS MVP delivery, security audits, AI agent builds, and browser QA.
中文简介正在准备中。当前显示原文说明。
Master REST and GraphQL API design principles to build intuitive, scalable, and maintainable APIs that delight developers. Use when designing new APIs, reviewing API specifications, or establishing API design standards.
中文简介正在准备中。当前显示原文说明。
Manage API versioning and evolution with URL/header/query strategies, deprecation workflows, breaking change classification, sunset headers, and consumer-driven contract testing. Use when designing versioning strategy, deprecating endpoints, or evolving API contracts.
中文简介正在准备中。当前显示原文说明。
Analyzes codebase structure, data flow, module relationships, and key patterns to generate and maintain a living architecture document (ARCHITECTURE.md).
中文简介正在准备中。当前显示原文说明。