Context Hub CLI
The chub CLI gives an agent curated, versioned API docs by language, keeps local notes and takes ratings; a skill and an MCP server are included
- 13,986 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The program uses the network: it downloads documents from a CDN and by default sends telemetry and ratings
- The skill lets the agent install a global npm package if `chub` is missing
- Document notes enter the agent's context and may contain harmful instructions
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/context-hubDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The program uses the network: it downloads documents from a CDN and by default sends telemetry and ratings
- The skill lets the agent install a global npm package if `chub` is missing
- Document notes enter the agent's context and may contain harmful instructions
This is third-party code. Review the repository files before installing.
What it does
Context Hub is a project with a `chub` command that hands an agent documentation for third-party APIs and libraries instead of answers from training data. The documents live in the repository itself as DOC.md files with a YAML header, split by version and language. The agent searches with `chub search "stripe payments"`, gets an id such as `stripe/api` and downloads the text with `chub get stripe/api --lang js`. The `--file` and `--full` flags let it take just the reference file it needs instead of the whole document. The `chub annotate` command saves a local note on a document, for example that webhook verification needs the raw request body. On the next `chub get` the program says a note exists and includes its text only with `--with-annotations`, because notes are treated as untrusted data. `chub feedback stripe/api up` sends a rating to the document's authors. For the agent to use the program on its own, the package ships a `get-api-docs` skill. It tells the agent to check `chub --help` before working with an external API, install `chub` if it is missing, find a document with `chub search`, fetch it with `--lang` and, after the task, save any gotchas it found as a note. The same package provides a `chub-mcp` binary: an MCP server with five tools, `chub_search`, `chub_get`, `chub_list`, `chub_annotate` and `chub_feedback`.
Who it is for. For developers whose agents mix up API methods and parameters, and for library authors who want to give agents their own documentation.
Good fit when
- The agent writes code for OpenAI, Stripe or another API and gets parameters wrong because it knows an old version
- You want a reference in one language, Python or JavaScript, without extra text for the others
- You want the agent to record the gotchas it finds and pull them in on later sessions
- A library author is ready to write the docs as DOC.md and submit them to the repository
- You need a local MCP server that searches documentation via `chub-mcp`
Not a fit when
- You work without network access: documents are fetched from a CDN and the machine only caches them
- Telemetry must not leave the machine and you cannot change settings: it is on by default
- You need docs for a closed or internal API: the repository holds material for public services, your own documents you have to write and connect yourself
- You need docs for any library at the version in your project: Context7 searches libraries, while here only what exists in `content` is available
Example request
Use chub to fetch the current OpenAI docs for Python and write a chat completions request with a streamed responseLimitations
Telemetry is on by default. Through PostHog the program records command usage, search queries, document popularity and errors, and it sends ratings separately to `api.aichub.org`. An anonymous client id is kept in `~/.chub/client_id`. Both are turned off with the variables `CHUB_TELEMETRY=0` and `CHUB_FEEDBACK=0` or with `telemetry: false` and `feedback: false` in `~/.chub/config.yaml`.
Documents are fetched from `cdn.aichub.org`, and the source can be changed in the configuration. The default cache refresh interval is 21600 seconds. The catalog so far holds only the documentation contributed to the repository; other content types, such as skills, are described by the authors as plans.
Note text lands in the agent's context, so notes included with `--with-annotations` count as untrusted input: the agent must not run commands written in them.
The `@aisuite/chub` package is at version 0.1.4, and the last change in the repository is dated 31 May 2026. Node.js 18 or newer is required. The license is MIT.
How to disable. Remove the program with `npm uninstall -g @aisuite/chub` and the copied `get-api-docs` skill file from the agent's skills folder. The cache, notes and client id live in `~/.chub`, and you can delete that folder. To keep the program but stop sending data, set `CHUB_TELEMETRY=0` and `CHUB_FEEDBACK=0`.
FAQ
Do I have to run chub myself?
No, the program is built for the agent. Ask it to use `chub` in a request or install the `get-api-docs` skill, and the agent calls the commands itself.
How do I install the skill in Claude Code?
Copy `SKILL.md` from the installed package to `.claude/skills/get-api-docs.md` in the project or to `~/.claude/skills/get-api-docs.md` for all projects. For Cursor the file goes to `.cursor/rules/get-api-docs.md`.
What do notes hold and how do they come back?
A note on a document is written locally with `chub annotate`. Later `chub get` only reports that a note exists, and you include its text with `--with-annotations`. `chub annotate --list` shows all notes and `chub annotate <id> --clear` removes them.
What do document authors get from ratings?
`chub feedback <id> up` or `down` sends a vote to a server where authors see which documents work. Sending is turned off with `CHUB_FEEDBACK=0`.
How many documents does Context Hub have?
The repository's `content` folder holds more than 1700 DOC.md files; versions and language variants of one document count as separate files. The repository tree also has 11 SKILL.md files, counting a test one and the package skill.
Discussion
No questions or reviews yet.
Related
CLI toolsUp-to-date, version-specific library docs and code examples in your agent's context, via MCP or a CLI plus skill
clihub turns any MCP server into a ready CLI: each tool becomes a subcommand and the result is one static binary with no runtime dependencies
An MCP server in one binary: AI reads and edits files across several projects and runs your commands
Ref's MCP server: documentation search for APIs and libraries that hands the agent only the relevant parts of pages
Context Engine: code search and memory for an agent
Context Engine
A skill that teaches an agent to query the Context Engine code index and symbol graph before reading files directly
Docs MCP Server
Grounded Docs MCP Server
A local MCP server that indexes docs from websites, GitHub, npm, PyPI and files and gives the agent version-aware search
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Context Hub (this entry) | no | 13,986 | ||
| Context7 | yes | API key | 63,873 | an MCP server that puts current library documentation of the right version into the agent's context |
| Ref MCP | yes | API key | 1,185 | an MCP server that searches API and library documentation and hands the agent only the relevant page fragments |
| Apple Doc MCP | no | not required | 632 | an MCP server that lets the agent search symbols and read Apple's |
README in short
The root has the project description with a quick start, contribution rules, a security policy and a code of conduct. The `cli` folder holds the JavaScript source: commands for search, fetch, notes, ratings and updates, BM25 search, a cache, the MCP server, and tests with fixture documents. The `content` folder stores the documents themselves, and GitHub Actions run checks and publish the content to a CDN.
README badge
Original SKILL.md
--- name: get-api-docs description: > Use this skill to get documentation for third-party APIs, SDKs or libraries before writing code that uses them to ensure you have the latest, most accurate documentation. This is a better way to find documentation than doing web search. This includes when a user asks for tasks like "use the OpenAI API", "call the Stripe API", "use the Anthropic SDK", "query Pinecone", or any other time the user asks you to write code against an external service and you need current API reference. --- # Get API Docs via chub When you need documentation for a library or API, fetch it with the `chub` CLI rather than guessing from training data. This gives you the current, correct API. ## Step 1 - Ensure chub exists and get the latest instructions for using it Run ```bash chub --help ``` ## Step 2 — Find the right documents ```bash chub search "<keywords>" --json ``` ## Step 3 — Fetch the docs ```bash chub get <id> --lang py # or --lang js, --lang ts ```