claude-code-sub-agents
Claude Code Subagents Collection
A collection of specialized subagents for Claude Code: frontend, backend, data, infrastructure, testing, security
Medium risk
We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.
Why this level
- The agents act on the project: write code, change files, run tests
- A clone into the global directory can conflict with other agents
Install
Manual install
mkdir ~/.claude/agents/lst97
cp /path/to/agents/*.md ~/.claude/agents/lst97Recommended way: keep the agents in a separate subfolder so you do not overwrite others. Substitute your own path to agents.
This is third-party code. Review the repository files before installing.
What it does
The repository adds a set of subagents to Claude Code, each responsible for its own area: coding in several languages, backend design, database work, infrastructure and deployment, review and testing, security. The agents are markdown files in the agents folder, grouped by category. Claude Code picks a fitting agent from task context or you invoke one explicitly. A separate organizer agent distributes work across the others in complex scenarios.
Who it is for. For developers who want a ready set of roles in Claude Code across the whole development cycle.
Good fit when
- You want a task handled by a narrowly specialized agent
- You want Claude Code to pick a role from context
- You need an organizer agent for a multi-step task
Not a fit when
- You do not work in Claude Code
- One command or skill is enough without a set of roles
Example request
Call backend-architect and design the API schema for the orders serviceLimitations
The set targets Claude Code and the global agents directory, so it can conflict with agents from other authors you already installed. The catalog is built around the author's personal full-stack workflow, and the role lineup reflects that.
How to disable. Remove the agents folder from ~/.claude/agents or the subfolder you copied them into.
Security check
- The agents act on the project: write code, change files, run tests
- A clone into the global directory can conflict with other agents
README in short
The README describes a collection of specialized subagents for Claude Code, grouped into categories: development, infrastructure, testing, data and AI, security, specialization. Installation copies the markdown files you need into ~/.claude/agents or clones the whole repository into ~/.claude. The author recommends a separate subfolder to avoid conflicts with agents already installed. An organizer agent distributes work. MIT licensed.
FAQ
How are agents selected automatically?
Claude Code reads the task context and picks a fitting agent, or you call one by name.
Can I install only some agents?
Yes, copy the markdown files you need from the agents folder into your agents directory instead of cloning everything.
Related
Open-source personal AI assistant on your own machine: answers in Telegram, Slack, Discord and WhatsApp, extended with skills and plugins
A self-improving agent from Nous Research with a TUI, messaging gateway, cron jobs and skills it writes itself
An open source coding agent for the terminal and desktop with build and plan modes
Open prompt library with a Claude Code plugin, MCP server and CLI: search, fetch and improve prompts and skills from an agent