Claude Code Subagents plugin
140 ready subagents for Claude Code: experts in languages, frameworks, databases, DevOps, testing and security.
- 1,024 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The files contain only subagent instructions without their own runtime code.
Install
- Works in
Manual install
git clone https://github.com/0xfurai/claude-code-subagents.gitRun it in `~/.claude`. Then keep the needed files in `~/.claude/agents/` and restart Claude Code.
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The files contain only subagent instructions without their own runtime code.
This is third-party code. Review the repository files before installing.
What it does
The repository holds 140 Claude Code subagent files in the `agents/` folder, one expert per technology: `python-expert`, `react-expert`, `rust-expert`, `nextjs-expert`, `fastapi-expert`, `ansible-expert` and others. Each file starts with a `name`, `description` and `model` block, then describes focus areas, approach, a quality checklist and output format. The README sections are languages and frameworks, web and frontend, games, databases, ORMs, infrastructure and DevOps, services, messaging, testing, data science and machine learning, monitoring, security, build tools, schema migrations, background jobs and package managers. The folder is cloned into `~/.claude`, after which Claude Code hands a task to the matching subagent by its description. A subagent can also be called explicitly: "Use the python-expert to optimize this algorithm".
Who it is for. Claude Code users who need ready role-specific subagents for different stacks without writing their own descriptions from scratch.
Good fit when
- One repository holds Python, React and Terraform and each part needs its own expert.
- A subagent for a specific framework such as FastAPI, Rails or Phoenix is needed as a starting template.
- You write your own subagents from a pattern: focus areas, approach, checklist and output format.
- Tasks are delegated automatically by description without choosing a role by hand.
Not a fit when
- You need one set of project rules: the subagents are generic and tied to none of your conventions.
- Context is limited and 140 subagent descriptions bloat it if all are installed at once.
- You use an agent without Claude Code subagent support: the file format is built for it.
Example request
Use fastapi-expert to review the service routes and suggest a dependency refactor without changing the public API.Limitations
The subagents are generic and do not know your conventions, so their output needs review like any model advice. If all 140 files go into `~/.claude/agents/`, their descriptions enter the context of every request, so keep only the ones you need. Files carry a `model` field with a fixed model, and if it is not available on your plan the field must be edited. Installation is described as cloning into `~/.claude`, and the working location is the `~/.claude/agents/` directory. The license is MIT, the last commit dates from November 2026.
How to disable. Delete the needed files from `~/.claude/agents/` or the whole clone folder. Subagents are picked up at Claude Code start, so restart the session after deleting.
FAQ
How do I install the subagents?
Run `cd ~/.claude` and `git clone https://github.com/0xfurai/claude-code-subagents.git`. Needed files must sit in `~/.claude/agents/`.
How do I call a subagent manually?
Name it in the request: "Use the python-expert to optimize this algorithm" or "Get the react-expert to refactor this component".
How does Claude Code pick a subagent?
By the `description` field at the top of the file: Claude Code delegates to the one whose description fits the context.
How do I add my own subagent?
Create an `.md` file in `agents/` with a lowercase hyphenated name, a description, focus areas, approach, checklist and output format.
How many subagents are in the collection?
The `agents/` folder has 140 files, and the README describes the collection as 100 or more specialized subagents.
Discussion
No questions or reviews yet.
Related
PluginsClaude Code Skills and Agents
Claude Code Skills & Agents
A Claude Code plugin bundling skills and sub-agents: deep research, image generation, GitHub work, code review and model switching
Claude Code kit with eight phase agents, 21 skills and hooks that keep the agent inside set limits.
A set of slash commands, skills and subagents for Claude Code: review, tests, security, releases, docs and GitHub to Linear sync
A set of Claude Code utilities: background skills, expert agents, slash commands and prompt templates for development
A ready-made Claude Code setup: agents, slash commands, skills, safety hooks, rules and MCP connections in one install
A Claude Code plugin: 39 commands for PRs, review, commits and specs, nine expert roles and hooks, in eight language versions
README in short
A collection of subagent description files for Claude Code under the MIT license. Each subagent describes its area, approach and response format.