Compile least-privilege policy from what your AI agents actually did, and hand over evidence the recipient verifies themselves.
- ✓Actively maintained (<30d)
- ✓Clear description
- ✓Documented (README)
- !No standard license detected
- !Install pipes a remote script into a shell (curl | sh)
git clone https://github.com/suhui-organization/ratchet{
"mcpServers": {
"ratchet": {
"command": "ratchet"
}
}
}MCP Servers overview
# Ratchet > **从 agent 的真实行为编译最小权限策略,并给出收货方能自己验证的证据。** ```bash curl -fsSL https://github.com/suhui-organization/ratchet/releases/latest/download/install.sh | sh ``` Linux / macOS(amd64 · arm64)。装完先跑这一条——**只读配置,不执行任何东西**: ```bash ratchet scan --home ~ ``` 它会在十秒内告诉你:这台机器上有几个 MCP server,其中几个**没锁版本**。 (我自己的机器:16 个里 12 个没锁。) 棘轮只能往一个方向转:权限可以收紧,放宽必须显式。 ## 这是什么 一个跑在**你自己机器上**的工具:它读 agent 用到的工具清单, 编译出一份最小权限策略,再把这次审计的产物打成一份 **收货方不需要信任你、也不需要装任何东西就能验证**的交付目录。 ## 这不是什么 - 不是沙箱、不是容器运行时; - 不是 prompt 内容过滤器; - 不是企业控制台(不做 dashboard、不做多租户、不做 SaaS 依赖); - 不要求你把日志传到哪里去。 ## 为什么值得存在 市场上的产品要么**让你自己手写策略**(Cedar / CEL / 自然语言), 要么**要你相信它的日志**。Ratchet 做的是这两件之间缺失的一步: **策略由观测编译,证据由收货方自己算哈希。** ## 一个例子 ```bash # 1) 编译:从工具清单得到一份保守的最小权限策略 ratchet policy draft --from examples/inventory.json --out policy.json # 2) 出交付物:报告 + 逐文件 sha256 清单 ratchet-report build --dir ./delivery --policy policy.json # 3) 收货方验证(不需要装 ratchet) python3 verify.py ./delivery ``` ## 仓库结构 | 目录 | 语言 | 是什么 | |---|---|---| | `cmd/` `internal/` | **Go** | 机器上的引擎:工具清单 → 能力判定 → 最小权限策略。编译成单个静态二进制,目标机器不需要任何运行时 | | `service/` | **Python** | 交付物侧:合规映射、sha256 清单、以及**只用标准库**的独立验证器 | | `web/` | **Nuxt 4 + Vue 3 + Tailwind** | 公开页面:落地页 + 收货方在浏览器里自己算哈希的验证页 | | `docs/` | — | 设计与计划(`DESIGN.md` / `PLAN.md` / `DECISIONS.md`) | ## 状态 **0.1.0 — 走的通的最小骨架**:策略编译与验证链路已实现且有测试, 真实环境扫描(读 harness 配置、MCP introspection)尚未实现,输入是清单文件。 详见 [docs/PLAN.md](docs/PLAN.md)。 ## 开发 ```bash make test # 三端全部测试 make go-test make py-test make web-test ```
What people ask about ratchet
What is suhui-organization/ratchet?
+
suhui-organization/ratchet is mcp servers for the Claude AI ecosystem. Compile least-privilege policy from what your AI agents actually did, and hand over evidence the recipient verifies themselves. It has 0 GitHub stars and its last recorded update is dated 2026-09-19.
How do I install ratchet?
+
You can install ratchet by cloning the repository (https://github.com/suhui-organization/ratchet) or following the README instructions on GitHub. ClaudeWave also provides quick install blocks on this page.
Is suhui-organization/ratchet safe to use?
+
Our security agent has analyzed suhui-organization/ratchet and assigned a Trust Score of 54/100 (tier: OK). See the full breakdown of passed checks and flags on this page.
Who maintains suhui-organization/ratchet?
+
suhui-organization/ratchet is maintained by suhui-organization. The last recorded GitHub activity is dated 2026-09-19, with 0 open issues.
Are there alternatives to ratchet?
+
Yes. On ClaudeWave you can browse similar mcp servers at /categories/mcp, sorted by popularity or recent activity.
Deploy ratchet to your cloud
Ship this repo to production in minutes. Each platform spins up its own environment with editable env vars.
Maintain this repo? Add a badge to your README
Drop the badge into your GitHub README to show it's tracked on ClaudeWave. Each badge links back to this page and reflects the live Trust Score.
[](https://claudewave.com/repo/suhui-organization-ratchet)<a href="https://claudewave.com/repo/suhui-organization-ratchet"><img src="https://claudewave.com/api/badge/suhui-organization-ratchet" alt="Featured on ClaudeWave: suhui-organization/ratchet" width="320" height="64" /></a>More MCP Servers
Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
An open-source AI agent that brings the power of Gemini directly into your terminal.
Real-time global intelligence dashboard. AI-powered news aggregation, geopolitical monitoring, and infrastructure tracking in a unified situational awareness interface
🕷️ An adaptive Web Scraping framework that handles everything from a single request to a full-scale crawl! Don't be shy, join here: https://discord.gg/EMgGbDceNQ
The fastest path to AI-powered full stack observability, even for lean teams.