github-actions-audit

github-actions-audit

Audits GitHub Actions workflow files for supply-chain risks like script injection, leaked tokens, unpinned actions, and broad permissions.

Category
访问服务器

README

GitHub Actions Security Audit

MCP server that audits .github/workflows/*.yml files for supply-chain risks. Catches script injection, leaked tokens, unpinned actions, broad permissions, and pull_request_target foot-guns — the patterns behind several 2024–2025 supply-chain incidents.

Built by Unbearable Labs. Pay-per-event — only billed when a tool is actually called.


Available on

  • Apify Actor Store — primary, metered usage (PPE)
  • MCPize — pending submission
  • MCP.so — pending submission
  • PulseMCP — pending submission
  • Smithery — pending submission
  • Glama — pending submission

Newsletter: Unbearable TechTips Weekly · All Actors: github.com/UnbearableDev

What it does

Point any MCP-capable client (Claude Desktop, Cursor, n8n, Make, Zapier, custom agents) at this server, hand it a workflow YAML, and get back structured findings with:

  • Severity — critical / high / medium / low / info
  • Affected job and step — exact location of the problem
  • Description — why it matters, with the actual attack vector
  • Remediation — what to do about it
  • Fix snippet — YAML you can paste directly

Tools

Tool Purpose
audit_workflow(workflow_yaml? | workflow_url?, min_severity='low') Run all checks
check_secrets(...) Secret-leakage paths only
check_permissions(...) GITHUB_TOKEN scope issues only
check_action_pinning(...) Action version-pinning only
check_runner_security(...) Self-hosted runner + script injection
check_workflow_config(...) Timeout / config hygiene
check_supply_chain_advanced(...) TeamPCP-class supply-chain patterns (GHA-201..208)
list_checks(category?) Browse the catalog

Provide exactly one of workflow_yaml (paste the content) or workflow_url (HTTPS URL — typically a GitHub raw URL to a specific workflow file).

Check catalog (v2: 21 checks)

ID Category Severity Title
GHA-001 secrets high Secret interpolated directly into run: script
GHA-002 secrets high Secret printed via echo / set-output
GHA-003 secrets medium Secret used in if: condition
GHA-004 secrets high Hardcoded credential pattern in env:
GHA-010 permissions high permissions: write-all granted
GHA-011 permissions medium No top-level permissions: (inherits broad default)
GHA-013 permissions high pull_request_target + checkout PR head = PWNing pattern
GHA-020 action_pinning high Third-party action pinned to mutable tag
GHA-021 action_pinning high Third-party action pinned to mutable branch
GHA-022 action_pinning medium First-party action not SHA-pinned
GHA-030 runner_security medium Self-hosted runner used on pull_request from forks
GHA-032 runner_security high Script injection via untrusted github.event.* interpolation
GHA-040 workflow_config low No timeout-minutes on job
GHA-201 supply_chain_advanced high Action pinned to unpinned branch ref (TeamPCP-class: @main/@master)
GHA-202 supply_chain_advanced high Action pinned to mutable tag — SHA pin recommended
GHA-203 supply_chain_advanced critical pull_request_target + checkout of PR head SHA/ref (codecov/tj-actions exploitation path)
GHA-204 supply_chain_advanced high Script injection via github.event.* user-controlled field in run:
GHA-205 supply_chain_advanced medium Action from non-allowlisted owner (untrusted 3rd-party)
GHA-206 supply_chain_advanced high Top-level permissions: write-all or contents: write without per-job scoping
GHA-207 supply_chain_advanced medium Secret logged via echo / cat in run: block
GHA-208 supply_chain_advanced low Action uses a known-retired tag

Pricing

Event USD
Any audit / check_* tool call $0.02
list_checks discovery $0.005

Connecting from Claude Desktop

{
  "mcpServers": {
    "gha-audit": {
      "transport": "streamable-http",
      "url": "https://YOUR-ACTOR-URL.apify.actor/mcp"
    }
  }
}

Sibling MCPs from Unbearable Labs

What's NOT covered (yet)

  • Reusable workflow auditing (multi-file resolution)
  • CodeQL-grade dataflow tracking
  • Marketplace-listed action reputation scoring

Source / contact

Source: github.com/UnbearableDev/github-actions-audit. Issues + ideas: unbearabledev@gmail.com.


📬 Built by Noel @ Unbearable Labs. More MCP servers + audit tips in the newsletter: https://unbearabletechtips.beehiiv.com

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选