A command-line tool for discovering, installing, and managing AI resources (commands, skills, agents, packages) across multiple AI coding tools including Claude Code, OpenCode, GitHub Copilot, and Windsurf.
- π¦ Centralized Repository: Manage all AI resources in one place
- π Symlink Installation: Install resources without duplication
- π€ Multi-Tool Support: Works with Claude Code, OpenCode, GitHub Copilot, and Windsurf
- π GitHub Integration: Import resources directly from GitHub repositories
- π― Pattern Matching: Install multiple resources using glob patterns
- β‘ Workspace Caching: Git repositories cached for 10-50x faster operations
- β Format Validation: Automatic validation of resource formats
- ποΈ Package Support: Group related resources together
aimgr supports four major AI coding tools:
| Tool | Commands | Skills | Agents | Directory |
|---|---|---|---|---|
| Claude Code | β | β | β | .claude/ |
| OpenCode | β | β | β | .opencode/ |
| VSCode / GitHub Copilot | β* | β | β * | .github/skills/, .github/agents/ |
| Windsurf | β | β | β | .windsurf/skills/ |
Notes:
- The support matrix reflects current aimgr direct-install support, not every upstream customization surface a tool may expose
- VSCode / GitHub Copilot and Windsurf support direct skill installation via aimgr
- Skills for Copilot and Windsurf use the same
SKILL.mdformat as other tools - Use
--target copilotor--target vscodefor GitHub Copilot installs (both names work) - Use
--target windsurffor Windsurf installs - GitHub Copilot / VS Code agent installs use
.github/agents/*.agent.md(installed artifact naming) - Repository source agents remain logical aimgr resources in
agents/*.md - GitHub Copilot / VS Code prompt files use
.github/prompts/*.prompt.md, but aimgr intentionally does not map/installcommandresources there - GitHub Copilot CLI has its own plugin/customization model for commands and slash commands, which is not the same as project-level
commands/*.mdinstalls
Install the latest release with a bootstrap script:
curl -fsSL https://raw.githubusercontent.com/dynatrace-oss/ai-config-manager/main/scripts/install.sh | shpowershell -NoProfile -ExecutionPolicy Bypass -Command "irm https://raw.githubusercontent.com/dynatrace-oss/ai-config-manager/main/scripts/install.ps1 | iex"Use AIMGR_VERSION to pin a release or AIMGR_INSTALL_DIR to override the install location.
If you already have Go installed, you can also install aimgr with:
go install github.com/dynatrace-oss/ai-config-manager/v3/cmd/aimgr@latest# Clone the repository
git clone https://github.com/dynatrace-oss/ai-config-manager.git
cd ai-config-manager
# Build and install to the OS-specific path from `make os-info`
make install
# Or just build (outputs to ./aimgr)
make buildPrecompiled binaries are published on the GitHub Releases page.
# 1. Configure default tools
aimgr config set install.targets claude
# 2. Add resources from GitHub
aimgr repo add gh:owner/repo
# 3. Add resources from local directory
aimgr repo add local:./my-resources
# 4. List available resources
aimgr repo list
# 5. Install resources in a project
cd your-project/
aimgr install skill/pdf-processing
aimgr install command/test agent/code-reviewer
# 6. Install multiple resources with patterns
aimgr install "skill/pdf*"Complete documentation is available in the docs/ directory:
- Getting Started - Installation, setup, and common workflows
- Configuration - Config file, environment variables, field mappings
- Sources - Managing local and GitHub sources (
ai.repo.yaml) - Repairing Resources - Fix broken installations, clean up projects
- Pattern Matching - Glob patterns for batch operations
- Output Formats - JSON, YAML, table output for scripting
- Resource Validation - Validate skills, agents, commands, and packages before sync/install
- Supported Tools - Tool compatibility and external documentation links
- Troubleshooting - Common issues and solutions
- Repository Layout - Internal folder structure
- Workspace Caching - Git repository caching for performance
- Git Tracking - Git-backed repository with change history
- Contributing Guide - How to contribute to aimgr
- Repository Overview - High-level architecture and repo map
- Change Workflow - Commit, push, branch, PR, and merge expectations
- Architecture - System design and components
- Code Style - Coding standards and conventions
- Testing - Test guidelines and patterns
- Development Environment - Setup for development
- Release Process - Creating releases
- AGENTS.md - Quick reference for AI coding agents
Contributions are welcome! See CONTRIBUTING.md for development setup, code style, and how to submit changes.
This project is licensed under the MIT License - see the LICENSE file for details.