zeno-validating-work
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseValidating Work (Contributor Workflow)
工作验证(贡献者工作流程)
Overview
概述
Run verification commands and confirm actual results before claiming work is complete or submitting. Evidence before assertions.
Trigger: About to commit, push, or claim task complete
Action: Run tests/verification → Confirm output → Then submit
在宣告工作完成或提交之前,运行验证命令并确认实际结果。先有证据,再做断言。
触发时机: 即将提交代码、推送代码或宣告任务完成时
操作步骤: 运行测试/验证 → 确认输出结果 → 再提交
The Gate Function
关卡函数
BEFORE claiming completion or submitting:
1. IDENTIFY: What proves this task is done?
- Tests? Which command?
- Linter? Which command?
- Manual verification? What to check?
2. RUN: Execute the FULL command fresh
- Don't trust previous runs
- Run it now in this session
3. READ: Full output, check exit code, count failures
4. VERIFY: Does output confirm success?
- YES: State completion WITH evidence
- NO: Fix issues, re-run, or report actual status
5. ONLY THEN: Commit, push, or claim complete
Skip any step = unverified claim在宣告完成或提交之前:
1. 明确:什么能证明任务已完成?
- 测试?用哪个命令?
- 代码检查工具?用哪个命令?
- 手动验证?需要检查什么?
2. 执行:完整运行命令(全新执行)
- 不要信任之前的运行结果
- 立即在当前会话中运行
3. 查看:完整输出内容,检查退出码,统计失败数量
4. 验证:输出结果是否确认成功?
- 是:附上证据宣告完成
- 否:修复问题、重新运行或报告实际状态
5. 仅在此时:提交代码、推送代码或宣告完成
跳过任何步骤 = 未经验证的宣告Verification Commands by Component
各组件的验证命令
stdb (Rust):
bash
cd stdb && cargo test && cargo clippy -- -D warningsstdb(Rust):
bash
cd stdb && cargo test && cargo clippy -- -D warningsExpected: All tests pass, no warnings
预期结果:所有测试通过,无警告
**backend (Deno):**
```bash
cd backend && deno task test && deno task lint
**backend(Deno):**
```bash
cd backend && deno task test && deno task lintExpected: All tests pass, no lint errors
预期结果:所有测试通过,无代码检查错误
**frontend (Node):**
```bash
cd frontend && npm run lint:all && npm run typecheck
**frontend(Node):**
```bash
cd frontend && npm run lint:all && npm run typecheckExpected: No lint errors, no type errors
预期结果:无代码检查错误,无类型错误
**docs / text contributions:**
```bash
**文档/文字贡献:**
```bashContent quality self-check before submitting:
提交前的内容质量自查:
| Check | Pass Means |
|---|---|
| Task complete | Every item in the task description is addressed |
| No filler | No "In this document we will..." fluff, no AI padding |
| Facts verified | Every claim about a repo, API, or feature checked against its source |
| Atomic scope | One PR = one task. No bundling unrelated work |
| Evidence | Command output, screenshots, or test results included |
| Readable | No obvious spelling errors, coherent structure |
undefined| 检查项 | 通过标准 |
|---|---|
| 任务完成度 | 任务描述中的每一项都已处理 |
| 无冗余内容 | 没有“在本文档中我们将...”这类空话,无AI生成的凑字数内容 |
| 事实已验证 | 所有关于仓库、API或功能的声明均已对照来源核实 |
| 原子性范围 | 一个PR对应一个任务。不捆绑无关工作 |
| 有证据 | 包含命令输出、截图或测试结果 |
| 可读性 | 无明显拼写错误,结构连贯 |
undefinedCommon Verification Needs
常见验证需求
| Task Type | Verify | Command Example |
|---|---|---|
| Feature | Tests pass | |
| Bug fix | Regression test + original symptom fixed | Run reproduction steps |
| Refactor | No behavioral changes | Full test suite |
| Docs | Links/formatting | Manual check or link checker |
| 任务类型 | 验证内容 | 命令示例 |
|---|---|---|
| 功能开发 | 测试通过 | |
| Bug修复 | 回归测试通过 + 原始问题已修复 | 运行复现步骤 |
| 代码重构 | 无行为变更 | 完整测试套件 |
| 文档 | 链接/格式正确 | 手动检查或使用链接检查工具 |
Anti-Patterns
反模式
❌ Wrong: "Should work now" without running tests
✅ Right:
[Run tests] [See: 12/12 pass] "Tests pass, ready to submit"❌ Wrong: "Linter passed" for compile check
✅ Right: Separate verification: then
cargo checkcargo clippy❌ Wrong: Claiming completion while tests failing
✅ Right: Report actual status: "3 tests failing, need to fix X"
❌ Wrong: Trusting a previous test run
✅ Right: Fresh verification in current session
❌ 错误做法: 未运行测试就说“现在应该可以工作了”
✅ 正确做法:
[运行测试] [查看结果:12/12通过] "测试通过,可提交"❌ 错误做法: 用“代码检查工具已通过”代替编译检查
✅ 正确做法: 分开验证: 然后
cargo checkcargo clippy❌ 错误做法: 测试失败时仍宣告完成
✅ 正确做法: 报告实际状态:“3个测试失败,需要修复X问题”
❌ 错误做法: 信任之前的测试运行结果
✅ 正确做法: 在当前会话中重新验证
Red Flags - STOP
危险信号 - 立即停止
- About to push without running verification
- Using "should", "probably", "seems" before verification
- Expressing satisfaction before checking: "Done!" (without verifying)
- Thinking "just this once" skip verification
- 未运行验证就准备推送代码
- 验证前使用“应该”“可能”“似乎”这类表述
- 未检查就表示满意:“完成了!”(未验证)
- 想着“就这一次”跳过验证
Integration
流程衔接
Before: (completing the work)
After: (push and create PR)
executing-taskssubmitting-work之前: (完成工作)
之后: (推送代码并创建PR)
executing-taskssubmitting-workBottom Line
核心原则
Run verification. Read output. Then claim completion.
No shortcuts. No assumptions. Evidence first.
先运行验证,查看输出结果,再宣告完成。
没有捷径,不要假设。证据优先。