verify-goal
Skill119 starsChanged 54 days agoMITView on GitHub
Independently verifies whether a completed task meets its defined goal conditions. Implements maker/checker separation — the agent that wrote the code does not grade its own work. Use after task execution to validate completion before marking done.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- k1ein-chen/Harness-Starterword for word, first seen Sep 24, 2026
Other versions1 project uses a version of this.
Word for word · 1
- k1ein-chen/Harness-Starter★ 119100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository8 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- handoverSkill
- harness-gcSkill
- harness-initSkill
- harness-modeSkill
- harness-startSkill
- tech-reviewSkill
The file
Files
Verify Goal
核心理念
写代码的 Agent 不能给自己打分。
一个任务执行完毕时,所谓的"完成了"只是执行者的声称。verify-goal 作为独立的验证者,以完成条件为基准逐条验证,输出 PASS / FAIL / PARTIAL 结果。
触发条件
在以下场景中必须触发:
/goal循环的每一轮结束后- 多步骤任务执行完毕后,声称"完成了"之前
- 提交 PR 前的最终检查
不触发:
- 用户只是问问题,不需要执行
- 简单的文件操作(重命名、移动、删除)
工作流
Step 1:获取目标定义
从以下来源获取完成条件(按优先级):
- 本次任务中用户或 AI 显式定义的完成条件(推荐格式见 CLAUDE.md Goal-Driven Execution)
- 对话上下文中可推导的隐含完成条件
- 任务本身的性质(如"修 bug" → 定义是"bug 复现步骤不再触发")
Step 2:逐条验证
对每条完成条件,执行对应的验证命令并记录结果:
| 条件类型 | 验证方式 | 示例 |
| 测试通过 | 运行测试命令 | npm test -- --testPathPattern=auth |
| 类型检查 | 运行类型检查 | tsc --noEmit |
| Lint | 运行 linter | npm run lint |
| 构建 | 运行 build | npm run build |
| 边界条件 | 检查文件是否被误改 | git diff --name-only 核对范围 |
| 自定义 | 执行用户定义的条件 | 按实际情况 |
Step 3:输出验证报告
使用以下固定格式输出:
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔎 目标完成验证报告
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
目标:[本次任务的目标描述]
验证结果:✅ 全部通过 / ⚠️ 部分通过 / ❌ 未通过
逐项验证:
1. [条件 A] ✅ PASS — [命令] → [输出摘要]
2. [条件 B] ❌ FAIL — [命令] → [失败摘要]
3. [条件 C] ✅ PASS — [命令] → [输出摘要]
─────────────────────────────────────
[全部通过]
结论:目标已完成,可进入下一步。
[部分通过 / 未通过]
失败项:
- [条件 B] → [失败原因]
建议:返回执行 Agent 修复后重新验证。
降级:连续 3 次验证未通过 → 停止循环,汇报给人类。
─────────────────────────────────────Step 4:判断下一步
根据验证结果决定后续动作:
| 验证结果 | 后续动作 |
| ✅ 全部通过 | 标记完成,更新状态文件,通知主流程继续 |
| ⚠️ 部分通过 | 将失败项返回执行 Agent 修复,修复后重新验证 |
| ❌ 全部未通过 | 停止循环,汇报给人类 |
| 连续 3 次 FAIL | Circuit Breaker 触发 — 停止循环,等待人类介入 |
与相关 skill 的关系
tech-review(事前) verify-goal(事后) acceptance(终验)
方案查完了没有? → 做完了条件满足了没? → 整个流程走完了没?
执行前触发 每轮 / 每次任务后触发 项目初始化结束时触发三者形成完整链条:事前查方案 → 事后验结果 → 终验看全貌。
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

