← 返回
未分类 中文

Yeeth Claw

Supply chain security hooks for Claude Code. Intercepts npm, pip, yarn, pnpm, and cargo install commands before execution and checks each package for supply...
用于 Claude Code 的供应链安全钩子,拦截 npm、pip、yarn、pnpm 和 cargo 安装命令并在执行前检查每个软件包的供应链安全。
bkojusner
未分类 clawhub v1.0.0 1 版本 99705.9 Key: 无需
★ 0
Stars
📥 339
下载
💾 0
安装
1
版本
#latest

概述

yeeth-claw

Claude Code hooks for supply chain security. Intercepts package installation

commands and flags suspicious packages before Claude executes them.

Hooks

OpenClaw

PreToolUse hook that watches npm install, pip install, yarn add,

pnpm add, and cargo add commands. For each package it checks:

  1. Package age — packages published < 30 days ago are flagged; < 7 days triggers a block
  2. Typosquat detection — Levenshtein distance against ~100 high-value targets per ecosystem (npm, PyPI, crates.io)
  3. Install scripts — postinstall hooks on flagged packages are noted as an additional risk signal

Risk tiers:

TierConditionExit code
---------
WARNAge < 30d or typosquat score ≥ 0.651 (non-blocking)
BLOCKAge < 7d and typosquat hit, or score ≥ 0.852 (blocks install)
ARGUSAny BLOCK + Argus API configured2 + submits for full analysis

Installation

git clone https://github.com/yeeth-security/yeeth-claw.git
cd yeeth-claw/hooks/openclaw
bash install.sh

The install script copies the hook to ~/.claude/hooks/openclaw/ and merges

the Claude Code settings if jq is available.

Restart Claude Code after installation for the hook to take effect.

Argus Integration

To enable full package analysis via the Argus API:

export OPENCLAW_ARGUS_URL=https://app.yeethsecurity.com
export OPENCLAW_ARGUS_KEY=<your-api-key>

When both are set, any BLOCK-tier package is submitted to Argus for full

static analysis and the job URL is included in the block message.

Requirements

  • Python 3.8+
  • No third-party dependencies (stdlib only)

Claude Code settings snippet

{
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "Bash",
        "hooks": [
          {
            "type": "command",
            "command": "python3 ~/.claude/hooks/openclaw/hook.py"
          }
        ]
      }
    ]
  }
}

版本历史

共 1 个版本

  • v1.0.0 当前
    2026-05-07 13:18 安全 安全

安全检测

腾讯云安全 (Keen)

安全,无风险
查看报告

腾讯云安全 (Sanbu)

安全,无风险
查看报告

🔗 相关推荐

ai-agent

self-improving agent

pskoett
捕获经验教训、错误及修正内容,以实现持续改进。适用于以下场景:(1)命令或操作意外失败;(2)用户纠正Claude(如“不,那不对……”“实际上……”);(3)用户请求的功能不存在;(4)外部API或工具出现故障;(5)Claude发现自身
★ 4,082 📥 812,424
ai-agent

Skill Vetter

spclaudehome
AI智能体技能安全预审工具。安装ClawdHub、GitHub等来源技能前,检查风险信号、权限范围及可疑模式。
★ 1,228 📥 268,039
dev-programming

Github

steipete
使用 `gh` CLI 与 GitHub 交互,通过 `gh issue`、`gh pr`、`gh run` 和 `gh api` 管理议题、PR、CI 运行及高级查询。
★ 676 📥 325,626