GitHub - ampedweb/awesome-subagents: A collection of 100+ specialized subagents covering a wide range of development use cases, originally just for Claude Code, but now with a script to convert them to Cursor and Opencode compatible syntax. · GitHub
Skip to content

ampedweb/awesome-subagents

 
 

Repository files navigation

Awesome Subagents

A curated collection of AI coding agent definitions for Claude Code, OpenCode, and Cursor.

Awesome Agent Count Last Update

Forked from VoltAgent/awesome-claude-code-subagents and extended to support multiple AI coding tools.


Overview

This repository contains 131+ specialised AI agent definitions organised into 10 categories. The definitions in categories/ use the Claude Code format as the canonical source. The generate.sh script produces equivalent definitions for OpenCode and Cursor.

Cursor also reads .claude/agents/ and ~/.claude/agents/ natively for compatibility, but the generated Cursor definitions in agent-specific/cursor/ add readonly: true support for agents that have no write, edit, or bash tools.

Agent Storage Locations

Tool Global Project
Claude Code ~/.claude/agents/ .claude/agents/
OpenCode ~/.config/opencode/agents/ .opencode/agents/
Cursor ~/.cursor/agents/ .cursor/agents/

Installation

Claude Code

Claude Code reads the definitions in categories/ directly. No generation step is needed.

Option 1 - Symlink via setup.sh (recommended):

git clone https://github.com/ampedweb/awesome-subagents.git
cd awesome-subagents
./setup.sh global                        # global access across all projects
./setup.sh project /path/to/your-project # project-level only

The script creates a single symlink named awesome-subagents/ inside each tool's agents directory, so agents appear as awesome-subagents/01-core-development/backend-developer, etc.

Option 2 - Copy manually:

git clone https://github.com/VoltAgent/awesome-claude-code-subagents.git
cd awesome-claude-code-subagents
./install-agents.sh

OpenCode

OpenCode requires generated definitions. Run generate.sh first, then setup.sh.

git clone https://github.com/ampedweb/awesome-subagents.git
cd awesome-subagents
./generate.sh                            # generates agent-specific/opencode/ and agent-specific/cursor/
./setup.sh global                        # symlinks OpenCode into ~/.config/opencode/agents/
./setup.sh project /path/to/your-project # symlinks OpenCode into the project

Cursor

Cursor requires generated definitions for full readonly support. Run generate.sh first, then setup.sh.

git clone https://github.com/ampedweb/awesome-subagents.git
cd awesome-subagents
./generate.sh                            # generates agent-specific/cursor/ (with readonly: true where applicable)
./setup.sh global                        # symlinks Cursor into ~/.cursor/agents/
./setup.sh project /path/to/your-project # symlinks Cursor into the project

Cursor also reads .claude/agents/ natively, so installing for Claude Code covers basic Cursor support. The generated Cursor definitions additionally apply readonly: true to agents that have no write, edit, or bash tools, and map the model field to Cursor-native model IDs (sonnet -> claude-4.6-sonnet, opus -> claude-4.6-opus, haiku -> claude-4.5-haiku).

Scripts Reference

generate.sh - Translates Claude Code definitions into OpenCode and Cursor formats.

Command Description
./generate.sh Generate all definitions (default)
./generate.sh clean Remove all generated files
./generate.sh list List source agents with details
./generate.sh help Show help

setup.sh - Creates or removes symlinks between agent directories and tool config. All commands present an interactive checkbox selector to choose which tools (Claude Code, OpenCode, Cursor) to link or unlink.

Command Description
./setup.sh global Symlink into global config
./setup.sh project <path> Symlink into a project's local config
./setup.sh unlink global Remove global symlinks
./setup.sh unlink project <path> Remove project symlinks

Categories

Plugin: voltagent-core-dev

Essential development subagents for everyday coding tasks.

Plugin: voltagent-lang

Language-specific experts with deep framework knowledge.

Plugin: voltagent-infra

DevOps, cloud, and deployment specialists.

Plugin: voltagent-qa-sec

Testing, security, and code quality experts.

Plugin: voltagent-data-ai

Data engineering, ML, and AI specialists.

Plugin: voltagent-dev-exp

Tooling and developer productivity experts.

Plugin: voltagent-domains

Domain-specific technology experts.

Plugin: voltagent-biz

Product management and business analysis.

Plugin: voltagent-meta

Agent coordination and meta-programming.

Plugin: voltagent-research

Research, search, and analysis specialists.

🤖 Understanding Subagents

Subagents are specialized AI assistants that enhance Claude Code's capabilities by providing task-specific expertise. They act as dedicated helpers that Claude Code can call upon when encountering particular types of work.

What Makes Subagents Special?

Independent Context Windows
Every subagent operates within its own isolated context space, preventing cross-contamination between different tasks and maintaining clarity in the primary conversation thread.

Domain-Specific Intelligence
Subagents come equipped with carefully crafted instructions tailored to their area of expertise, resulting in superior performance on specialized tasks.

Shared Across Projects
After creating a subagent, you can utilize it throughout various projects and distribute it among team members to ensure consistent development practices.

Granular Tool Permissions
You can configure each subagent with specific tool access rights, enabling fine-grained control over which capabilities are available for different task types.

Core Advantages

  • Memory Efficiency: Isolated contexts prevent the main conversation from becoming cluttered with task-specific details
  • Enhanced Accuracy: Specialized prompts and configurations lead to better results in specific domains
  • Workflow Consistency: Team-wide subagent sharing ensures uniform approaches to common tasks
  • Security Control: Tool access can be restricted based on subagent type and purpose

Getting Started with Subagents

1. Access the Subagent Manager

/agents

2. Create Your Subagent

  • Choose between project-specific or global subagents
  • Let Claude generate an initial version, then refine it to your needs
  • Provide detailed descriptions of the subagent's purpose and activation triggers
  • Configure tool access (leave empty to inherit all available tools)
  • Customize the system prompt using the built-in editor (press e)

3. Deploy and Utilize Your subagent becomes immediately available. Claude Code will automatically engage it when suitable, or you can explicitly request its help:

> Have the code-reviewer subagent analyze my latest commits

Subagent Storage Locations

Type Path Availability Precedence
Project Subagents .claude/agents/ Current project only Higher
Global Subagents ~/.claude/agents/ All projects Lower

Note: When naming conflicts occur, project-specific subagents override global ones.

📖 Subagent Structure

Each subagent follows a standardized template:

---
name: subagent-name
description: When this agent should be invoked
tools: Read, Write, Edit, Bash, Glob, Grep
model: sonnet
---

You are a [role description and expertise areas]...

[Agent-specific checklists, patterns, and guidelines]...

## Communication Protocol
Inter-agent communication specifications...

## Development Workflow
Structured implementation phases...

Tool Assignment Philosophy

Smart Model Routing

Each subagent includes a model field that automatically routes it to the right Claude model — balancing quality and cost:

Model When It's Used Examples
opus Deep reasoning — architecture reviews, security audits, financial logic security-auditor, architect-reviewer, fintech-engineer
sonnet Everyday coding — writing, debugging, refactoring python-pro, backend-developer, devops-engineer
haiku Quick tasks — docs, search, dependency checks documentation-engineer, seo-specialist, build-engineer

You can override any agent's model by editing the model field in its frontmatter. Set model: inherit to use whatever model your main conversation is using.

Tool Assignment Philosophy

Each subagent's tools field specifies Claude Code built-in tools, optimized for their role:

  • Read-only agents (reviewers, auditors): Read, Grep, Glob - analyze without modifying
  • Research agents (analysts, researchers): Read, Grep, Glob, WebFetch, WebSearch - gather information
  • Code writers (developers, engineers): Read, Write, Edit, Bash, Glob, Grep - create and execute
  • Documentation agents (writers, documenters): Read, Write, Edit, Glob, Grep, WebFetch, WebSearch - document with research

Each agent has minimal necessary permissions. You can extend agents by adding MCP servers or external tools to the tools field.

🧰 Tools

Claude Code skill for browsing and fetching subagents from this catalog.

Command Description
/subagent-catalog:search <query> Find agents by name, description, or category
/subagent-catalog:fetch <name> Get full agent definition
/subagent-catalog:list Browse all categories
/subagent-catalog:invalidate Refresh cache

Installation:

cp -r tools/subagent-catalog ~/.claude/commands/

Contributions are welcome. See CONTRIBUTING.md for guidelines.

  • Add new agent definitions via pull request
  • Improve existing definitions
  • Report issues

Licence

MIT - see LICENSE

All agent definitions are provided "as is" without warranty. The maintainers do not audit or guarantee the security or correctness of any definition and accept no liability for issues arising from their use.

If you find an issue with a listed agent or want your contribution removed, please open an issue.

About

A collection of 100+ specialized subagents covering a wide range of development use cases, originally just for Claude Code, but now with a script to convert them to Cursor and Opencode compatible syntax.

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

Contributors

Languages

  • Shell 100.0%