context-hub
Local library for Claude Code and Codex skills, prompts, MCP servers, plugins, and environment variables. Desktop app, CLI, and local web UI; delivers each asset into the agent's own workspace or global directory.
Install / Use
claude mcp add bb-boy680 -- npx -y github:bb-boy680/context-hubIf the server publishes to npm under a different name, use that package instead โ check the repo README.
MCP Server
Model Context Protocol server
Quality Score
Category
Development & EngineeringSupported Platforms
Tags
Our assessment of context-hub
context-hub scores 83/100 on our quality scale, 3028th of 4,658 Development & Engineering skills we index.
Its MCP Server is 7.5 KB long, well organised into 14 sections with 4 code examples: a thorough specification that gives an agent plenty to work with.
It has 10 GitHub stars, so there is little community track record yet; judge it on its content.
Maintenance, license and trust
- The repository was last updated 2 days ago, so context-hub is actively maintained.
- It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 97/100, with no cautions. These come from repository metadata, not a code audit โ read the skill file before letting an agent act on it.
context-hub compared with similar skills
All 4 of these similar skills score higher than context-hub; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| context-hub (this skill)by bb-boy680 | 83 | 10 | 2d ago | MCP Server |
| Agent-Reachby Panniantong | 100 | 106.6k | 5d ago | CLAUDE.md |
| taste-skillby Leonxlnx | 100 | 106.0k | today | CLAUDE.md |
| headroomby headroomlabs-ai | 100 | 78.3k | today | CLAUDE.md |
| ai-job-searchby ai-job-search-dev | 100 | 67.6k | 3d ago | CLAUDE.md |
Frequently asked questions
- How do I install context-hub?
- Run
claude mcp add bb-boy680 -- npx -y github:bb-boy680/context-hub. The install tabs above show the steps for each supported agent. - Which AI agents does context-hub work with?
- It is written for Claude Code, Claude Desktop and OpenAI Codex, as a MCP Server file. Other agents that read the same format can often use it too.
- Is context-hub safe to use?
- It is MIT-licensed and scores 97/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is context-hub still maintained?
- The repository was last updated 2 days ago, so context-hub is actively maintained.
Skill content
View source on GitHubContext Hub
One local hub for everything your AI coding agents read.
Context Hub keeps skills, MCP servers, plugins, prompts, and environment variables in one place, then delivers each one into the agent's own workspace or global directory. It runs as a desktop app, a standalone CLI, or a local web page.
Install: desktop builds and CLI binaries ยท from source
<img src="assets/screenshot-skills.png" alt="Context Hub skills library" width="820" /> </div>Why Context Hub?
Every agent keeps its own assets in its own places: skills in a workspace folder, MCP servers in a JSON config, prompts in a commands directory, variables in a shell profile. Add a second agent and the same asset has to be maintained twice.
Context Hub gives those assets one catalog and one install flow, while still leaving every file where the agent expects to find it.
Screenshots
| MCP servers | Extensions |
| ----------------------------------------- | ----------------------------------------------- |
|
|
|
Features
- One library for every asset: Skills, MCP servers, plugins, prompts, and environment variables in a single local catalog.
- Agent-native delivery: The Claude Code and Codex integrations, installed from the extensions marketplace, deliver each asset into that agent's own workspace or global directory; more agents arrive as extensions.
- Declare once, use everywhere: MCP servers and prompts are written into each agent's own config. MCP servers can be enabled or disabled per client; prompts are installed or removed.
- Desktop, CLI, or browser: One local hub powers the Electron app, standalone binaries, and the built web page.
- Extensible by design: Agent integrations, asset modules, search providers, settings sections, and language packs are all plugins.
- Global search: Press
Ctrl/Cmd+Kto search every module from anywhere. - Keeps itself current: Skills and plugin packs imported from a source track their upstream; the desktop app updates itself from GitHub Releases.
- Local-first and bilingual: Your assets stay in the agents' own directories, and the UI ships in English and Simplified Chinese.
Install
Desktop app
Download the latest build from Releases:
| Platform | File |
| --------------------- | --------------------------------------------- |
| Windows x64 | Context-Hub-<version>-win-x64.exe |
| macOS (Apple Silicon) | Context-Hub-<version>-mac-arm64.dmg |
| Linux x64 | Context-Hub-<version>-linux-x86_64.AppImage |
[!NOTE] The builds are not code-signed. Windows SmartScreen and macOS Gatekeeper will warn on first launch. The macOS build is currently arm64 only.
Standalone CLI
Each binary embeds the hub and the page, so the CLI itself does not need
Node.js. Installing a skill from a package registry uses npx, so that one path
needs Node.js on the machine:
| Platform | File |
| --------------------------- | ---------------------------------------------------- |
| Windows x64 | context-hub-windows-x64.exe |
| Linux x64 / arm64 | context-hub-linux-x64, context-hub-linux-arm64 |
| macOS Apple Silicon / Intel | context-hub-darwin-arm64, context-hub-darwin-x64 |
Run it and open the URL it prints:
./context-hub-linux-x64
# context hub server listening on http://127.0.0.1:...
The CLI accepts --port, --host, --folder, --home, and --static; run it
with --help for the full list.
From source
Node.js 20 or newer and pnpm 10.33.4 are required.
pnpm install
pnpm dev # web app and desktop app together
pnpm dev:web # the browser form only
pnpm dev:desktop # the desktop shell only
Usage
- Choose your agents in the sidebar. Only the agents you select are read and written.
- Add a workspace: the folder of the project you are working on. The hub reads each agent's workspace directories inside it.
- Add assets: import a skill from a folder or a registry your plugins provide, add an MCP server, install a plugin pack, write a prompt, or record an environment variable.
- Install to a place: skills, MCP servers, plugins, and prompts can be delivered to a workspace or a global agent directory.
[!TIP] Press
Ctrl/Cmd+Kto search skills, servers, packs, prompts, and variables without leaving the page you are on.The hub keeps its library in
~/.context-hub. SetCONTEXT_HUB_HOMEto use a different directory.
Supported agents
- Claude Code: install the integration from the Context Hub extensions marketplace; the integration declares which assets it accepts and where it writes them.
- Codex: install the integration from the same marketplace, under the same contract.
- More agents: arrive as marketplace extensions; Cursor and others are not done.
Development
pnpm dev # web + desktop, with HMR
pnpm build # build every package (page + desktop bundles)
pnpm --filter web start # serve the built form
pnpm build:cli # build the CLI for five targets (needs bun)
pnpm build:desktop # bundle the desktop main and preload processes
pnpm test # every package's tests in one process
pnpm lint # ESLint
pnpm typecheck # tsc across the workspace
Desktop installers are built per platform:
pnpm --filter @context-hub/desktop package:win
pnpm --filter @context-hub/desktop package:mac
pnpm --filter @context-hub/desktop package:linux
The test suite writes real files, starts real HTTP listeners, and spawns a real
MCP server; one update test also clones a git repository, so git must be on
PATH.
The kernel is built on Cordis; the UI uses React 19, Vite, Tailwind CSS, and shadcn/ui.
Documentation
- Building the distributable forms: CLI, desktop, release versioning, and what can be verified locally.
- Extensions: package layout, market index, local development sources, and the trust model.
- Releasing: the release workflow, in Chinese with screenshots.
License
MIT. See LICENSE.
If this is useful, a star helps others find it.
Related Skills
Agent-Reach
106.6kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu โ one CLI, zero API fees.
taste-skill
106.0kAgent skills that give coding agents taste: art direction, reference images and a browser check instead of generic AI pages
headroom
78.3kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
ai-job-search
67.6kThe job search that runs on your machine. AI job application framework built on Claude Code: evaluate postings, tailor CVs, write cover letters, prep interviews. Fork it and own it.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit โ see the Safety scan above for what the skill file itself contains.
