Verification Before Completion
- 914 installs
- 7.3k repo stars
- Updated July 23, 2026
- jnmetacode/superpowers-zh
verification-before-completion is a Claude Code skill that requires fresh command output before any completion, bug-fix, or test-pass claim for developers who must prove work with evidence before merging or shipping.
About
verification-before-completion is a Claude Code skill (version 1.0.0, MIT) that blocks unsubstantiated success claims during development. Its iron rule: no fresh verification evidence means no completion claim. A five-step gate function requires identifying the proof command, running it fully, reading complete output and exit codes, verifying the conclusion, and only then stating results. A failure-mode table maps common claims—tests pass, linter clean, build succeeded, bug fixed, requirements met—to required evidence versus insufficient substitutes like prior runs or assumptions. Developers reach for verification-before-completion before commits, PRs, or handoffs when agents or teammates might skip rerunning tests.
- 5-step gate function that must run before any claim of success
- Iron rule: no fresh verification evidence = no permission to declare done
- Explicit mapping of 7 common conclusions versus what actually qualifies as proof
- Prevents all forms of rationalization and proxy trust in agent reports
- Hard-gate before commit, PR creation, or final handoff to next skill
Verification Before Completion by the numbers
- 914 all-time installs (skills.sh)
- +45 installs in the week ending Jul 28, 2026 (Skillselion tracking)
- Ranked #141 of 1,382 Code Review & Quality skills by installs in the Skillselion catalog
- Security screen: LOW risk (skills.sh audit)
- Data as of Jul 28, 2026 (Skillselion catalog sync)
npx skills add https://github.com/jnmetacode/superpowers-zh --skill verification-before-completionAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 914 |
|---|---|
| repo stars | ★ 7.3k |
| Security audit | 3 / 3 scanners passed |
| Last updated | July 23, 2026 |
| Repository | jnmetacode/superpowers-zh ↗ |
How do you prove tests passed before merging?
Ensure every claim of completion, bug fix, or test pass is backed by fresh command output before merging or shipping.
Who is it for?
Developers and agent sessions that must substantiate test-pass, lint-clean, build-success, or bug-fix claims with rerunnable shell evidence before git commit or PR.
Skip if: Teams that only need a test-writing generator or static code review without a behavioral gate on completion claims.
When should I use this skill?
Agent or user is about to claim work is done, tests pass, linter is clean, or a bug is fixed without having run verification in the current turn.
What you get
Fresh command output with exit codes, failure counts, and evidence-backed completion statements.
- Evidence-backed completion statement
- Fresh command output with exit code
By the numbers
- Version 1.0.0 under MIT license
- Five-step verification gate function before any completion claim
Files
完成前验证
概述
在没有验证的情况下宣称工作完成，这不是高效，而是不诚实。
核心原则： 始终用证据支撑结论。
对这条规则敷衍了事，就等于违背了它的精神。
铁律
没有新鲜的验证证据，不许宣称完成如果你在这条消息中没有运行验证命令，就不能声称测试通过。
门控函数
在宣称任何状态或表达满意之前：
1. 确定：什么命令能证明这个结论？
2. 运行：执行完整命令（重新运行，完整执行）
3. 阅读：完整输出，检查退出码，统计失败数
4. 验证：输出是否支持这个结论？
- 如果否：用证据说明实际状态
- 如果是：带证据陈述结论
5. 只有这时：才能做出结论
跳过任何一步 = 说谎，不是验证常见失败模式
| 结论 | 需要 | 不够格 |
|---|---|---|
| 测试通过 | 测试命令输出：0 failures | 之前的运行结果、"应该会通过" |
| Linter 无报错 | Linter 输出：0 errors | 部分检查、推断 |
| 构建成功 | 构建命令：exit 0 | linter 通过、日志看起来没问题 |
| Bug 已修复 | 测试原始症状：通过 | 代码改了，假设已修复 |
| 回归测试有效 | 红-绿循环已验证 | 测试只通过了一次 |
| 代理已完成 | VCS diff 显示变更 | 代理报告"成功" |
| 需求已满足 | 逐项核对清单 | 测试通过 |
红线——停下来
- 使用"应该"、"大概"、"似乎"
- 验证前就表达满意（"太好了！"、"完美！"、"搞定！"等）
- 即将提交/推送/创建 PR 却没有验证
- 信任代理的成功报告
- 依赖部分验证
- 想着"就这一次"
- 累了想赶紧收工
- 任何暗示成功但实际未运行验证的措辞
防止合理化
| 借口 | 现实 |
|---|---|
| "应该能行了" | 运行验证命令 |
| "我有信心" | 信心 ≠ 证据 |
| "就这一次" | 没有例外 |
| "Linter 通过了" | Linter ≠ 编译器 |
| "代理说成功了" | 独立验证 |
| "我累了" | 疲劳 ≠ 借口 |
| "部分检查就够了" | 部分检查什么也证明不了 |
| "换个说法这条规则就不适用了" | 精神大于字面 |
关键模式
测试：
✅ [运行测试命令] [看到：34/34 pass] "全部测试通过"
❌ "应该能通过了" / "看起来对了"回归测试（TDD 红-绿）：
✅ 编写 → 运行（通过）→ 回退修复 → 运行（必须失败）→ 恢复 → 运行（通过）
❌ "我写了回归测试"（没有经过红-绿验证）构建：
✅ [运行构建] [看到：exit 0] "构建通过"
❌ "Linter 通过了"（linter 不检查编译）需求：
✅ 重读计划 → 创建核对清单 → 逐项验证 → 报告缺口或完成
❌ "测试通过了，阶段完成"代理委派：
✅ 代理报告成功 → 检查 VCS diff → 验证变更 → 报告实际状态
❌ 信任代理报告为什么这很重要
来自 24 次失败记录：
- 搭档说"我不信你"——信任被破坏
- 未定义的函数被交付——会直接崩溃
- 遗漏需求被交付——功能不完整
- 虚假完成浪费的时间 → 返工 → 重做
- 违反原则："诚实是核心价值。如果你说谎，就会被替换。"
何时使用
以下情况之前必须使用：
- 任何形式的成功/完成声明
- 任何满意的表达
- 任何关于工作状态的正面陈述
- 提交、创建 PR、标记任务完成
- 进入下一个任务
- 委派给代理
本规则适用于：
- 准确措辞
- 同义词和换一种说法
- 暗示成功
- 任何传达完成/正确性的沟通
底线
验证没有捷径。
运行命令。阅读输出。然后才能宣称结果。
这没有商量余地。
Related skills
Improve Codebase ArchitectureSafely deepen clusters of shallow modules into cohesive, testable units while respecting their external dependencies.531k185k
Caveman ReviewGet ultra-compressed, one-line code review comments that cut noise while keeping every actionable fix.260k92.5kCodebase DesignShared vocabulary for designing deep modules: improve a module's interface, find deepening opportunities, decide where a seam goes, make code more testable.233k185kCavecrewDelegate coding tasks to specialized subagents that return compressed output, keeping the main context window usable for much longer sessions.210k92.5k
Requesting Code ReviewDispatch a consistent, high-signal code reviewer subagent that catches plan deviations and quality issues before merging or continuing development.178k260kCode ReviewReviews a branch or PR diff on two axes at once: conformance to coding standards plus a code-smell baseline, and whether it actually implements the original spec.167k185kFAQ
What counts as proof that tests passed?
verification-before-completion requires fresh test command output showing 0 failures and a successful exit code in the current session. Prior runs, partial checks, or assumptions like "should pass" are explicitly insufficient.
When should verification-before-completion run?
verification-before-completion applies before claiming work complete, a bug fixed, or tests passing—especially immediately prior to git commit or PR creation when no verification command ran in the current message.
Is Verification Before Completion safe to install?
skills.sh reports 3 of 3 security scanners passed. Review the Security Audits panel on this page before installing in production.