kenryu42/cc-safety-net 预览 preview

kenryu42/cc-safety-net

A pre-execution guard for AI coding agents. It blocks destructive Git and file system commands, plus common attempts to access sensitive files, before a tool call runs. Supports Amp Code, Antigravity CLI, Claude Code, Codex, Cursor, DeepSeek Harness, Gemini CLI, GitHub Copilot CLI, Grok Build, Hermes Agent, Kimi Code, OpenClaw, OpenCode, and Pi.

Project Overview项目介绍

CC Safety Net (kenryu42/cc-safety-net) is a cross-platform safety plugin for coding-agent CLIs that intercepts destructive shell commands and access to secrets such as SSH keys and .env files before the tool call is executed. It semantically parses each command to determine real intent, so wrapping or reordering flags cannot bypass the policy, and a broken configuration file never causes a blanket block. The repository ships parallel integrations for Windows, macOS, and Linux across Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot CLI, OpenCode, Antigravity CLI, Grok Build, Kimi Code, Amp Code, Hermes Agent, DeepSeek Harness, and several more, with one install path documented per CLI on ccsafetynet.com.

In a typical workflow a developer installs the matching plugin for their preferred CLI, after which the analyzer runs synchronously ahead of every tool call; matched rules can reject, allow, or prompt the user, and every decision is appended to an audit log for later review. The /cc-safety-net slash command functions as an operator manual that explains why a command was blocked, helps draft or migrate custom rules, adjusts the active policy or safety level, and verifies whether secret protection is enabled in a given CLI like Codex. The skill is shipped with the Claude Code and Codex plugins and is built directly into the OpenCode and Pi integrations, and it only loads its context when the user explicitly invokes it.

The library surface is published on npm as cc-safety-net and exposes a checkCommand function callable from any Node.js program, which must receive an absolute cwd and whose thrown errors signal a hard block. The project is MIT-licensed, includes CI plus Codecov coverage, and the README explicitly warns that only the analyzer and a subset of Windows integrations are covered by automated tests, leaving the remaining Windows CLI paths as best-effort. First-time operators should read the Security Model and Known Limitations guides and configure recovery so a corrupted rulebook cannot silently disable protection.

CC Safety Net(kenryu42/cc-safety-net)是一款跨编码代理 CLI 的安全防护插件,在工具执行前阻断破坏性命令并拦截对 SSH 密钥、.env 等敏感文件的访问。它通过语义解析判断命令真实意图,命令包装或参数重排无法绕过,配置文件出错也不会造成全量封锁。项目在 Windows、macOS、Linux 上同时支持 Claude Code、Codex、Cursor、Gemini CLI、GitHub Copilot CLI、OpenCode、Antigravity CLI、Grok Build、Kimi Code、DeepSeek Harness 等代理。

典型使用流程:开发者按所使用 CLI 安装对应插件或集成层,激活后命令执行前自动经过分析引擎;命中策略可被拒绝、放行或询问,事件写入审计日志供事后溯源。/cc-safety-net 子命令作为操作手册,解释拦截原因、迁移自定义规则、调整安全等级或验证密钥保护是否启用。插件面向把多款编码代理混用于生产环境、担心自动化误删仓库或泄露密钥的工程团队与个人开发者。

依赖方面,库 API 通过 cc-safety-net npm 包发布,Node.js 环境下可直接调用 checkCommand;插件以 MIT 协议开源。文档明确指出 Windows 上仅分析器及部分集成经过自动化测试,其余 CLI 属于尽力兼容、未做完整覆盖。配置文件被设计为失败安全(fail-open),但建议首次部署前仍阅读 Security Model 与 Known Limitations 以确认边界。

Pre-install check安装前体检Compatibility · Security兼容性 · 安全性 no risk signal found未发现风险信号
  • This site's static screen found no obvious risk signal (stars, license, activity, manifest)本站静态筛查没发现明显风险信号(星标、许可证、更新活跃度、清单完整度)
DSH walks through these 9 checksDSH 会逐条核对这 9 项

Compatibility兼容性

  • DSH, Node, OS and profile requirementsDSH 版本 / Node 版本 / 操作系统 / profile 是否满足要求
  • External dependencies and runtimes (Electron / Python / Docker, ...)外部依赖与运行时(Electron / Python / Docker 等)是否齐备
  • Conflicts with installed plugins: command names, skill / tool names, ports, duplicate MCP registration与已装插件是否冲突:命令名、skill / tool 重名、端口占用、重复 MCP 注册

Security安全性

  • Repo matches the facts registered here; archived or abandoned?仓库是否与页面登记一致,是否归档或长期停更
  • Safety of preinstall / install / postinstall and install.sh / setup.ps1preinstall / install / postinstall 与 install.sh、setup.ps1 是否安全
  • curl|bash, download-then-execute, obfuscation, unrelated domains → stop immediatelycurl|bash、下载即执行、混淆代码、无关域名 → 立刻停止
  • Typosquatting or unmaintained packages among the new dependencies新增依赖里有没有 typosquatting 或无人维护的包
  • Requested permissions vs. what the feature actually needs申请了哪些权限、是否超出功能所需(filesystem / network / shell / clipboard)
  • Any sudo / admin requirement, plus uninstall and rollback是否要求 sudo / 管理员权限,以及卸载与回滚方式

Anything uncertain must be marked unknown with a note on how to confirm it. This site's signal screen is a static snapshot, not a security audit.拿不准的必须标「未知」并说明要我怎么确认。本站的信号筛查是静态快照,不能替代安全审计。

Or use CLI install (for developers)或使用命令行安装(适合开发者)

CLI Install命令行安装

dsh plugin --profile web add github:kenryu42/cc-safety-net

把 kenryu42/cc-safety-net 加入你的 DSH 配置(web profile)即可启用。

READMEREADME

CC Safety Net

CI codecov Version License: MIT

English · 简体中文 · 日本語

https://github.com/user-attachments/assets/928dbe97-31e3-41d1-b35a-7941a701b056

CC Safety Net (Coding CLI Safety Net) blocks destructive commands and access to secrets such as SSH keys and .env files before the tool call runs. It parses what the command does. Wrapping the command or reordering flags does not hide it. A broken config file never blocks anything. It is not a sandbox: it does not contain processes, set filesystem permissions, or watch network egress.

[!NOTE] Full documentation → covers installation, configuration, reference material, guides, and the security model. This README is the short version.

How it works

An AI coding agent tries to run a command or open a file. CC Safety Net checks what it would actually do before it runs. Safe calls such as git status run normally; dangerous ones such as git reset --hard never run, and the agent is told why.

Showing the opening section of the README — the full document lives in the repository以上为 README 开头摘要,完整文档在仓库内 · View the full README on GitHub →在 GitHub 查看完整 README →

← 上一个 Prev dsh-im 下一个 Next brooks-lint →