oss PR审查

AI 智能社区 v2.0.0

作者仓库星标 122
许可证 MIT
作者仓库 Happycapy-skills

解读按原文结构重写，命令、链接、术语均保留；右侧可核对作者原始 SKILL.md

方法与流程

Quick Start：原文以此为独立章节。

适用与边界

当前原文没有单列适用场景。
当前原文没有单列不适用场景或限制。

原文中的明确线索

要点：「9-agent swarm」、「> "I don't just write PRs.」、「This skill deploys a 9-agent swarm that autonomously: 1.」、「See README.md for full documentation.」
文件与命令：git、config/swarm-config.json、fixes/features、./scripts/run-swarm.sh、24/7、PRs/day、CONTRIBUTING.md、README.md

流狐整理：以上内容来自当前 SKILL.md 的章节与原词；未补写作者没有声明的工具、兼容性或能力。

OSS Contributor Swarm Skill

"I don't just write PRs. I continuously contribute to open source while learning from every interaction."

This skill deploys a 9-agent swarm that autonomously:

Qualifies repositories for contribution-friendliness
Finds suitable "good first issues"
Analyzes requirements
Clones and understands codebases
Writes fixes/features
Adds tests (if needed)
Updates documentation (if needed)
Creates professional pull requests
Responds to review feedback automatically

Quick Start

# Run a single contribution cycle
./scripts/run-swarm.sh single

# Run continuous 24/7 mode (target: 3-5 PRs/day)
./scripts/run-swarm.sh start

# View dashboard
./scripts/run-swarm.sh dashboard

Prerequisites

GitHub CLI (gh) - authenticated
Claude Code CLI (claude)
jq and git

Key Features

Continuous Mode

Runs 24/7, automatically finding and contributing to open source projects.

Learning System

Tracks success patterns and improves over time:

Best issue types (typo, docs, tests)
Best languages and repo sizes
Maintainer preferences
Auto-avoids problematic repos

Auto-Fix Reviews

Agent 8 can automatically fix simple review requests:

Typos, style, formatting → Auto-fix
Logic changes → Escalate to human

Configuration

Edit config/swarm-config.json to customize:

Target daily PRs (default: 3-5)
Max concurrent PRs (default: 3)
Repo qualification criteria
Learning parameters

Safety

Never force pushes
Respects CONTRIBUTING.md
Max 5 files / 100 lines per PR
Human approval for security changes

See README.md for full documentation.

流狐档案作者与许可取自来源；运行、权限和网络为流狐检测或估算

流狐分类: AI 智能
作者声明 Agent: 未找到明确声明；不据此推断已兼容或已测试
静态检查: 98 / 100 · 启发式扫描，不代表运行安全
作者 / 版本 / 许可: @happycapy-ai · v2.0.0 · MIT
流狐 Token 估算: 低消耗
流狐接入估算: 即装即用
是否需要外部 API Key: 未发现要求
检测到的系统要求: 未声明
底层运行要求: 未声明
检测到的文件与系统行为: 只读
允许写入 / 修改
检测到的网络行为: 仅限本地
安装命令数: 无（仅作为资料）

档案由构建时根据 SKILL.md 与安装命令自动衍生，可能与作者实际意图存在差异。

需要注意： 未限定 allowed-tools，默认拥有全部工具权限。

输出预览 oss-contributor-swarm.preview

作者没有在当前 SKILL.md 中定义固定输出样例。

SKILL.md SKILL.md

# OSS Contributor Swarm Skill

> "I don't just write PRs. I continuously contribute to open source while learning from every interaction."

This skill deploys a **9-agent swarm** that autonomously:
1. Qualifies repositories for contribution-friendliness
2. Finds suitable "good first issues"
3. Analyzes requirements
4. Clones and understands codebases
5. Writes fixes/features
6. Adds tests (if needed)
7. Updates documentation (if needed)
8. Creates professional pull requests
9. Responds to review feedback automatically

## Quick Start

```bash
# Run a single contribution cycle
./scripts/run-swarm.sh single

# Run continuous 24/7 mode (target: 3-5 PRs/day)
./scripts/run-swarm.sh start

# View dashboard
./scripts/run-swarm.sh dashboard
```

## Prerequisites

- GitHub CLI (`gh`) - authenticated
- Claude Code CLI (`claude`)
- `jq` and `git`

## Key Features

### Continuous Mode
Runs 24/7, automatically finding and contributing to open source projects.

### Learning System
Tracks success patterns and improves over time:
- Best issue types (typo, docs, tests)
- Best languages and repo sizes
- Maintainer preferences
- Auto-avoids problematic repos

### Auto-Fix Reviews
Agent 8 can automatically fix simple review requests:
- Typos, style, formatting → Auto-fix
- Logic changes → Escalate to human

## Configuration

Edit `config/swarm-config.json` to customize:
- Target daily PRs (default: 3-5)
- Max concurrent PRs (default: 3)
- Repo qualification criteria
- Learning parameters

## Safety

- Never force pushes
- Respects CONTRIBUTING.md
- Max 5 files / 100 lines per PR
- Human approval for security changes

See README.md for full documentation.

证据边界与执行链路

作者原文负责流程事实；流狐只索引当前章节、要点、文件与命令。

章节    -> Quick Start → Prerequisites → Key Features → Continuous Mode → Learning System → Auto-Fix Reviews
要点    -> 9-agent swarm · > "I don't just write PRs. · This skill deploys a 9-agent swarm that autonomously: 1. · See README.md for full documentation.
文件/命令 -> git · config/swarm-config.json · fixes/features · ./scripts/run-swarm.sh · 24/7 · PRs/day · CONTRIBUTING.md · README.md
内容 SHA-256 -> c72ef1028162