# Verify Before Done

> 宣称"做好了/修好了/测试过了/能跑"之前用，提交、建 PR 或 MR 之前也用。要求先真跑验证命令、看到输出再下结论——证据先于断言。

- **Type:** Skill
- **Install:** `agentstack add skill-chipfighter-coding-discipline-verify-before-done`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [chipfighter](https://agentstack.voostack.com/s/chipfighter)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [chipfighter](https://github.com/chipfighter)
- **Source:** https://github.com/chipfighter/coding-discipline/tree/main/plugins/coding-discipline/skills/verify-before-done

## Install

```sh
agentstack add skill-chipfighter-coding-discipline-verify-before-done
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

宣称做完之前，先拿出这条消息里刚跑出来的证据。没在本条消息里跑过的，就不能说它过。

## 闸门（下任何"成功/完成"结论前）
1. 这条结论靠哪个命令证明？
2. 把那条命令**完整、重新**跑一遍（别用上一次的结果）。
3. 读全部输出 + 退出码 + 数失败条数。
4. 输出对得上结论吗？对——结论连证据一起说；不对——按真实状态说。
5. 这一步走完，才能开口下结论。

跳过任何一步 = 撒谎，不是验证。

## 跑不了 ≠ 跳过
命令跑不起来（环境缺、依赖坏、超时）就是**失败**，按失败报。不许当"这步略过"，悄悄绿过去。

## 这些都不算验证过
- lint 过了 ≠ 能编译 / 能跑。
- 子 agent 报"成功" ≠ 真做了——自己去看 diff / 实际产物核对。
- 回归测试"跑一次过了" ≠ 有效——要红绿验证：先回退修复 → 必须挂 → 再装回 → 过。
- "需求达成"靠逐条对着需求勾，不是"测试绿了就算完"。
- 只跑了一部分 ≠ 全过。

## 红线
- 不准用"应该""大概""看起来对"代替证据。
- 没跑验证前不准庆祝（"好了！""搞定！""完美！"）。
- 改了要紧东西想直接 commit / push / 提 PR 或 MR 前，先过上面的闸门。

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [chipfighter](https://github.com/chipfighter)
- **Source:** [chipfighter/coding-discipline](https://github.com/chipfighter/coding-discipline)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-chipfighter-coding-discipline-verify-before-done
- Seller: https://agentstack.voostack.com/s/chipfighter
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.