Skip to content
Back to search

rtfmbro-mcp

rtfmbro provides always-up-to-date, version-specific package documentation as context for coding agents. An alternative to context7

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/marckrenn--rtfmbro-mcp/badge.svg)](/listings/marckrenn--rtfmbro-mcp)

HTML

<a href="/listings/marckrenn--rtfmbro-mcp"><img src="/api/listings/marckrenn--rtfmbro-mcp/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Quick Start

Claude Code

claude mcp add-json rtfmbro '{ "type": "http",  "url": "https://rtfmbro.smolosoft.dev/mcp/" }'

Claude Desktop / VS Code / etc.

Add the remote server to your MCP configuration:

{
  "rtfmbro": {
    "type": "http", 
    "url": "https://rtfmbro.smolosoft.dev/mcp/"
  }
}

Cursor

Agent Integration

To truly integrate rtfmbro with your AI coding agent of choice, copy the appropriate meta-instruction file to your project:

- **GitHub Copilot**: Copy [`.github/copilot-instructions.md`](.github/copilot-instructions.md) to your project - **Claude Code**: Copy [`CLAUDE.md`](CLAUDE.md) to your project root - **Cursor**: Copy [`.cursor/rules`](.cursor/rules) to your project root - **Other agents**: Adapt the instructions from either file above to your agent's format

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 154 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # rtfmbro-mcp
  2. > rtfmbro provides always-up-to-date, version-specific package documentation as context for coding agents. An alternative to [context7](https://github.com/upstash/context7).
  3. https://github.com/user-attachments/assets/dbe0b3b4-a42c-4e91-8bcd-a94d430ef0b8
  4. Demonstration of rtfmbro in use, fetching a specific version (3.1.1) of Flask's [readme](https://github.com/pallets/flask/blob/3.1.1/README.md) and [docs](https://github.com/pallets/flask/tree/3.1.1/docs).
  5. ## Overview
  6. rtfmbro is a Model Context Protocol (MCP) server that provides real-time, version-aware documentation fetching for packages across multiple ecosystems. It bridges the gap between AI agents and accurate, up-to-date package documentation by fetching docs directly from GitHub repositories at the exact version your project uses.
  7. ## Supported Ecosystems
  8. | Ecosystem | Registry | Status |
  9. |-----------|----------|--------|
  10. | **Python** | PyPI | ✅ Full Support |
  11. | **Node.js** | npm | ✅ Full Support |
  12. | **Swift** | SPM | 🚧 Alpha |
  13. | **GitHub** | Direct | ⚠️ Fallback |
  14. ## Why rtfmbro?
  15. ### The Problem
  16. - **Stale Documentation**: AI models often rely on outdated training data about packages
  17. - **Missing Context**: Source code in `node_modules` etc. lacks high-level documentation, browsing it is usually token-consuming and inefficient
  18. - **Version Mismatches**: Generic documentation doesn't match your specific package version, especially for legacy projects or brand new packages
  19. ### The Solution
  20. rtfmbro tries to solve these issues by:
  21. 1. **Version-Precise Fetching**: Retrieves documentation from the exact git tag/commit that matches your lockfile
  22. 2. **Comprehensive Coverage**: Extracts all documentation files (`.md`, `.mdx`, `.txt`, `.rst`, `.html`) from the repository
  23. 3. **Intelligent Caching**: SHA-based currency checking ensures docs stay fresh without unnecessary re-fetching
  24. 4. **Agent Integration**: Seamlessly integrates with AI coding assistants via the Model Context Protocol
  25. 5. **Great DX**: Zero seting up for developers, just add the server to your MCP configuration, instructions and start fetching docs
  26. ### MCP Tools
  27. The server exposes four primary tools to AI agents:
  28. | Tool | Purpose | Parameters | Returns |
  29. |------|---------|------------|---------|
  30. | `get_readme` | Fetches and returns the README file for a specific package version | `package`, `version`, `ecosystem` | README content as string |
  31. | `get_documentation_tree` | Generates a comprehensive folder structure of all documentation files | `package`, `version`, `ecosystem` | Tree structure as string |
  32. | `read_files` | Reads specific documentation files with optional line range slicing | `package`, `version`, `ecosystem`, `requests[]` | Dictionary mapping paths to content |
  33. | `search_github_repositories` | Searches for GitHub repositories using the GitHub Search API | `query`, `sort`, `order`, `per_page` | Formatted repository search results |
  34. ## Installation & Setup
  35. ### Quick Start
  36. #### Claude Code
  37. ```bash
  38. claude mcp add-json rtfmbro '{ "type": "http", "url": "https://rtfmbro.smolosoft.dev/mcp/" }'
  39. ```
  40. #### Claude Desktop / VS Code / etc.
  41. Add the remote server to your MCP configuration:
  42. ```json
  43. {
  44. "rtfmbro": {
  45. "type": "http",
  46. "url": "https://rtfmbro.smolosoft.dev/mcp/"
  47. }
  48. }
  49. ```
  50. #### Cursor
  51. [![Install MCP Server](https://cursor.com/deeplink/mcp-install-dark.svg)](https://cursor.com/install-mcp?name=rtfmbro&config=eyJ0eXBlIjoiaHR0cCIsInVybCI6Imh0dHBzOi8vcnRmbWJyby5zbW9sb3NvZnQuZGV2L21jcC8ifQ%3D%3D)
  52. ### Agent Integration
  53. To truly integrate rtfmbro with your AI coding agent of choice, copy the appropriate meta-instruction file to your project:
  54. - **GitHub Copilot**: Copy [`.github/copilot-instructions.md`](.github/copilot-instructions.md) to your project
  55. - **Claude Code**: Copy [`CLAUDE.md`](CLAUDE.md) to your project root
  56. - **Cursor**: Copy [`.cursor/rules`](.cursor/rules) to your project root
  57. - **Other agents**: Adapt the instructions from either file above to your agent's format
  58. ## How It Works
  59. ### Documentation Workflow
  60. 1. **Registry Lookup**: Queries the package registry (PyPI, npm, etc.) for metadata
  61. 2. **GitHub Discovery**: Extracts the GitHub repository URL from package metadata
  62. 3. **Version Resolution**: Fetches available git tags and matches them against your semantic version
  63. 4. **Smart Fetching**: Clones the repository at the exact matched tag/commit
  64. 5. **Content Filtering**: Extracts only documentation files, removing source code and build artifacts
  65. 6. **Caching & Currency**: Stores results with SHA-based currency checking for efficient re-access
  66. ### Caching Strategy
  67. - **SHA-Based Validation**: Compares current repository commit SHA with cached version
  68. - **Automatic Invalidation**: Re-fetches documentation when new commits are detected
  69. - **Persistent Storage**: Maintains local cache to avoid redundant GitHub API calls
  70. - **Metadata Preservation**: Stores documentation tree structure for fast browsing
  71. ## Prerequisites
  72. - Package must be published to a supported registry (PyPI, npm)
  73. - Package metadata must contain a valid GitHub repository link
  74. - Repository must use git tags for version management
  75. - Documentation files must be present in the repository (not just generated sites)
  76. ## Roadmap
  77. ### Near Term
  78. - [x] **Ecosystem independent fallback**: Implement a fallback mechanism for unsupported ecosystems
  79. - [ ] **Provide rtfmbro source code**: Open source the server codebase
  80. - [ ] **Public docker image**: Create a public Docker image for easy deployment
  81. - [ ] **Private repo support**: Allow authenticated access to private repositories
  82. - [ ] **Add Tests**: Implement unit and integration tests for core functionality
  83. - [ ] **Enhanced Python Support**: Include pydocs and docstring extraction
  84. - [ ] **Search Capabilities**: Search across documentation corpus
  85. ### Future Ecosystems / Languages / Registries
  86. - [ ] **Rust** ([crates.io](https://crates.io/))
  87. - [ ] **Go** ([pkg.go.dev](https://pkg.go.dev/))
  88. - [ ] **Java/Kotlin** ([Maven Central](https://central.sonatype.com/))
  89. - [ ] **C#/.NET** ([NuGet](https://www.nuget.org/))
  90. - [ ] **Ruby** ([RubyGems](https://rubygems.org/))
  91. ### Source code hosting and repository support:
  92. - [x] **GitHub**: Support for GitHub repositories
  93. - [ ] **Gitlab**: Support for GitLab repositories
  94. - [ ] **Bitbucket**: Support for Bitbucket repositories
  95. - [ ] **Launchpad**: Support for Launchpad repositories
  96. ## Known Issues
  97. - Some packages may have documentation in separate standalone repos
  98. - Large repositories may take a bit of time to clone and process initially
  99. ## Similar / Additive Projects
  100. - **[mcp-package-docs](https://github.com/sammcj/mcp-package-docs)**: Another MCP server for package documentation, focusing on documentation extraction, LSP servers, etc. May be a great supplement to rtfmbro.
  101. - **[rust-docs-mcp-server](https://github.com/Govcraft/rust-docs-mcp-server)**: MCP server for Rust documentation, focused on Rust-specific features and documentation formats.
  102. - **[mcp-ragdocs](https://github.com/qpd-v/mcp-ragdocs)**: MCP server for RAG (Retrieval-Augmented Generation) documentation, aimed at improving the documentation experience for AI models.
  103. - **[godoc-mcp](https://github.com/mrjoshuak/godoc-mcp)**: MCP server for Go documentation, providing access to Go package documentation via the Model Context Protocol.
  104. - **[context7](https://github.com/upstash/context7)**: Alternative to rtfmbro
  105. ### Differences between **context7** and **rtfmbro**
  106. |Aspect|context7|rtfmbro|
  107. |------|--------|-------|
  108. | **Actuality** | Scrapes documentation ahead-of-time at intransparent intervals or upon user trigger. As of writing, the "latest" Next.js docs are already 2 days old.| Fetches documentation just-in-time, ensuring it's always up-to-date. |
  109. | **Version-specific docs** | Theoretically allows scraping older versions (useful for legacy or longtime projects), but the process is complicated, limiting practical availability effectively to latest versions. | Fetches older documentation just-in-time, and always remains current, identical to latest docs. |
  110. | **Search strategy** | Uses either A) optionally token-limited RAG search to filter/preprocess docs which can be hit-or-miss, or B) dumps all content into LLM's context, resulting in excessive token-use. | Employs agentic discovery ([as used by Claude Code itself](https://x.com/pashmerepat/status/1926717705660375463)) |
  111. | **Developer Experience** | Requires explicitly mention of context7 in every prompt. | Operates via defined rules/instructions, auto-selecting appropriate package name/version from lock file ("set and forget"). |
  112. | **Support** | Language/ecosystem independent. | Currently language/ecosystem-specific; planned additional languages/ecosystems and language-independent fallback mechanism soon. |

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.