Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Leodorareluctant259/superpowers-zh --skill verification-before-completiongit clone --depth 1 https://github.com/Leodorareluctant259/superpowers-zhWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/leodorareluctant259/superpowers-zh/verification-before-completion)<a href="https://agentmods.dev/skills/leodorareluctant259/superpowers-zh/verification-before-completion"><img src="https://agentmods.dev/badge/skills/leodorareluctant259/superpowers-zh/verification-before-completion/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/leodorareluctant259/superpowers-zh/verification-before-completion"><img src="https://agentmods.dev/badge/skills/leodorareluctant259/superpowers-zh/verification-before-completion.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00050 | $0.01228 |
| Opus 5 | $0.00025 | $0.00614 |
| Sonnet 5 | $0.00010 | $0.00246 |
| Haiku 4.5 | $0.00005 | $0.00123 |
Grade A, and why
verification-before-completion scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
This is a copy
100% identical to verification-before-completion — 0 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
What it actually says
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。
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 140 lines · 50 tokens per session scan A 9e1152e80264
verification-before-completion is a skill published in the GitHub repository Leodorareluctant259/superpowers-zh (5 stars, last pushed yesterday), licensed MIT. It adds 50 tokens to every session and 1,228 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. It is 100% identical to verification-before-completion, differing in 0 lines, and is treated as a copy.
Other skills, from other repositories
intuitive-tests
Use this skill whenever the user asks about unit test best practices, test organization, flat test suites, redundant tests, test refactors, pytest/JUnit/Jest/xUnit layout, test taxonomy, flaky tests, coverage quality, fixtures, mocks, parametrization, pruning existing UTs, or "which tests are worth keeping." It…
agent-evaluation
Evaluate and improve Claude Code commands, skills, and agents. Use when testing prompt effectiveness, validating context engineering choices, or measuring improvement quality.
do-and-judge
Execute a task with sub-agent implementation and LLM-as-a-judge verification with automatic retry loop.
design-testing-strategy
Use before writing any type of tests. Distills 14 industry sources into deterministic decision gates, schemas, and worked test examples.
ainb-fleet:daemons
Runtime-health view of the four fleet daemons — phone bridge, notifyd, ATC, and the fleet auto-continue daemon — in one table. Each row reports state (running / stopped / unknown), pid, uptime, last activity, error count, and a HEALTH reason that distinguishes a clean stop from a crash (stale heartbeat) or a…
test-coverage
Use after writing tests to assess coverage quality across structural, mutation, requirements, and API/integration dimensions; organized knowledge for choosing and interpreting coverage analyses.