SkillAgentSearch skills...

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 1Better-Boy -- npx -y github:1Better-Boy/context-hub

If the server publishes to npm under a different name, use that package instead — check the repo README.

About this skill
🔌

MCP Server

Model Context Protocol server

Quality Score

83/100

Supported Platforms

Claude Code
Claude Desktop
OpenAI Codex

Our assessment of context-hub

context-hub scores 83/100 on our quality scale, 2983rd of 4,705 Development & Engineering skills we index.

Its MCP Server is 7.8 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.

Substance
29/30
Structure
20/20
Description
15/15
Adoption
4/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated 16 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 — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands.

Automated pattern scan on 2026-11-21. It catches known dangerous patterns, not every risk — read a skill 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.

SkillScoreStarsUpdatedFormat
context-hub (this skill)by 1Better-Boy831016d agoMCP Server
taste-skillby Leonxlnx100120.0ktodayCLAUDE.md
Agent-Reachby Panniantong100114.0ktodayCLAUDE.md
hyperframesby heygen-com10088.2ktodayCLAUDE.md
rufloby ruvnet10082.3ktodayCLAUDE.md

Frequently asked questions

How do I install context-hub?
Run claude mcp add 1Better-Boy -- npx -y github:1Better-Boy/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?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. 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 16 days ago, so context-hub is actively maintained.
<div align="center"> <img src="assets/icon.png" width="40" alt="Context Hub" />

Context 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 · from source

Skills MCP Prompts Plugins Desktop TypeScript Latest release

English · 简体中文

<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 | | ----------------------------------------- | ----------------------------------------------- | | 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+K to 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

  1. Choose your agents in the sidebar. Only the agents you select are read and written.
  2. Choose Global or add a workspace: Global shows only each agent's global locations. Selecting a project workspace includes both project and global locations.
  3. 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.
  4. Install to a place: skills, MCP servers, plugins, and prompts can be delivered to a workspace or a global agent directory. On each asset page, use Install locations to filter by one workspace or global location, or choose All to clear the filter. Existing skills in the shared workspace folder (.agents/skills) can still be browsed and collected there; the folder is no longer an install target.

[!TIP] Press Ctrl/Cmd+K to search skills, servers, packs, prompts, and variables without leaving the page you are on.

The hub keeps its library in ~/.context-hub. Set CONTEXT_HUB_HOME to 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

View on GitHub
GitHub Stars10
CategoryDevelopment
Updated16d ago
Forks0

Languages

TypeScript

Trust signals

97/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

1 info
context-hub — MCP Server: Install & Safety Check | SkillAgent