Token导航 LogoToken导航TokenDH.com
研究检索敏感数据clawhub未标认证来源可访问clear审计提醒

pre-publish-security预发布安全性

Agent Skill

用于辅助安全审计、权限检查、凭据风险、认证流程和常见漏洞排查。它适合让 Agent 梳理敏感配置、检查依赖风险、分析鉴权逻辑或生成安全复核清单。使用时不能把工具输出直接当最终结论,涉及密钥、令牌、用户数据或生产系统时,应先确认最小权限、脱敏方式和操作边界。

总安装

9,163

周安装

378

GitHub Stars

公开资料未说明

下载量

2,994
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:pre-publish-security(预发布安全性)
来源仓库:https://github.com/solmas/pre-publish-security
安装命令:
openclaw skills install pre-publish-security
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install pre-publish-security

简介

pre-publish-security 构建多层安全审核系统,防止 GitHub/ClawHub 发布时的凭证泄露与漏洞暴露。

  • 频率感知机制动态调整扫描强度,平衡安全性与开发效率。
  • 自动检测硬编码密钥、过时代码与文档敏感信息,生成修复建议报告。
  • 支持与 CI 系统集成,实现提交即扫描与阻断合并请求功能。
  • 输出结果需人工确认误报情况,并结合组织安全基线制定处置策略。

SKILL.md

name
pre-publish-security
description
Multi-layered security audit system for GitHub/ClawHub releases. Prevents credential leaks, detects vulnerabilities, validates documentation. Frequency-aware scanning (quick/history/dependencies). Blocks bad pushes automatically.
version
2.0.0
author
solmas
homepage
https://github.com/solmas/openclaw-pre-publish-security
license
MIT
tags
metadata
openclaw
requires
bins
[git, jq, grep]
optional
[npm, pip, safety, shellcheck]
install
user-invocable
true

Pre-Publish Security Protocol

Prevents security breaches like exposed credentials in open-source releases.

Features

Multi-Level Scanning

  • Quick scan: Every push (~5s)
  • History scan: Monthly deep dive (~2-5min)
  • Dependency CVE: Weekly npm/Python check (~30s)
  • Full audit: On-demand comprehensive (~3-6min)

Smart Frequency Management

  • State tracking knows when each scan last ran
  • Auto-determines which scans to run
  • Prevents redundant checks

What It Catches

  • GitHub PATs, API keys, passwords, private keys
  • Secrets in git history (even if "deleted")
  • npm/Python dependency CVEs
  • Unsafe code patterns (eval, exec)
  • Documentation placeholders ([ORG], example.com)
  • Missing LICENSE/README files
  • Exported environment variables with secrets

Automated Protection

  • Git pre-push hook blocks bad commits
  • Severity-based exit codes (CRITICAL/HIGH/MEDIUM/LOW)
  • Markdown reports with actionable fixes

Quick Start

Install Pre-Push Hook

# Automatic protection on every push
./install-hooks.sh /path/to/your/repo

Run First History Scan

# One-time deep dive (or monthly)
./audit-full.sh /path/to/repo history

Check Status

# See when scans last ran
./schedule.sh status

Run Scheduled Audits

# Auto-determines what to run based on time
./schedule.sh run /path/to/repo

Manual Scans

# Quick scan (every push)
./audit-simple.sh /path/to/repo

# Git history scan (monthly)
./audit-full.sh /path/to/repo history

# Dependency scan (weekly)
./audit-full.sh /path/to/repo dependencies

# Full audit (before releases)
./audit-full.sh /path/to/repo full

What Gets Scanned

Quick Scan (Every Push)

  • Current file secret patterns
  • Documentation placeholders
  • Basic license/README presence
  • Runtime: ~5 seconds

History Scan (Monthly)

  • Full git commit history
  • Deleted-but-accessible credentials
  • Historical security issues
  • Runtime: 2-5 minutes

Dependency Scan (Weekly)

  • npm audit (Node.js CVEs)
  • Python safety check
  • Known vulnerabilities
  • Runtime: ~30 seconds

Full Audit (On-Demand)

  • All of the above
  • Environment variable leaks
  • Pre-commit hook verification
  • Code quality patterns
  • Runtime: 3-6 minutes

Severity Levels

  • CRITICAL → Blocks push (secrets, credentials)
  • HIGH → Requires approval (vulnerabilities, missing LICENSE)
  • MEDIUM → Warning (TODOs, missing README)
  • LOW → Informational

Integration

Pre-Push Hook (Recommended)

./install-hooks.sh ~/my-repo
git push  # Automatic security check

Weekly Cron

# Add to OpenClaw cron
openclaw cron add \
  --name "weekly-repo-scan" \
  --cron "0 3 * * 1" \
  --announce \
  --message "Run: ~/.openclaw/workspace/skills/pre-publish-security/schedule.sh run ~/repo"

Manual Pre-Publish

# Before clawhub publish
./audit-full.sh ~/skills/my-skill full
clawhub publish skills/my-skill --version 1.0.1

Files

  • audit-simple.sh - Fast pre-push scan
  • audit-full.sh - Complete scanner with tracking
  • schedule.sh - Status & smart automation
  • install-hooks.sh - Git hook installer
  • audit-state.json - State tracking (auto-created)
  • AUDIT-SCHEDULE.md - Detailed frequency guide
  • README.md - Full documentation
  • agents/ - Sub-agent definitions (future use)

Requirements

Required:

  • git
  • jq
  • grep

Optional (enhanced detection):

  • npm (Node.js dependency scanning)
  • pip + safety (Python dependency scanning)
  • shellcheck (bash script validation)

State Tracking

Automatically tracks:

  • Last run timestamp for each scan type
  • Total scan counts
  • Cumulative findings by severity

View with: ./schedule.sh status

Exit Codes

  • 0 - Passed (no issues or low/medium only)
  • 1 - Critical issues (blocks push)
  • 2 - High issues (requires review)

Real-World Example

Problem: Accidentally pushed GitHub PAT in git remote URL Solution: This tool caught it and blocked the push Result: Credential never exposed publicly

Use Cases

  1. Individual Developers: Pre-push hook prevents accidents
  2. Open-Source Projects: Protects against contributor mistakes
  3. ClawHub Skills: Validates before publishing
  4. CI/CD: Add to GitHub Actions for automated checks
  5. Security Audits: Comprehensive repository review

Why This Exists

On 2026-03-15, a GitHub PAT was accidentally exposed in a git config file. This protocol ensures it never happens again - to anyone.

License

MIT - Use it, improve it, share it.

Contributing

Issues & PRs welcome at: https://github.com/solmas/pre-publish-security

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

75.69%
按下载量换算2,266

安全审计

VirusTotal

通过

ClawScan

可疑

Static analysis

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills