Manage complex projects with a team of AI assistants, smoothly and efficiently.
Agentic Project Management (APM) is a AI workflow framework that brings real-world project management principles into your AI-assisted workflows. It addresses a fundamental challenge of LLMs: context window limitations.
APM uses various context retention techniques, coordinating a team of specialized AI agents in a structured way so that you can maintain productive AI-assisted work for longer periods before facing model hallucinations and needing to start over. When context window does fill up, APM ensures a smooth transition to a "fresh" chat session without important context loss.
Think of it like having a project manager, developers, ad-hoc specialists, and a setup/configuration expert all powered by AI and working together under your guidance.
Accessing APM Assets: You have a few options to get the APM prompts, guides and protocol definitions:
-
Use the APM Template (Recommended for Custom Projects):
- Click the "Use this template" button on the APM GitHub Repository.
- This creates your own repository pre-filled with the entire APM structure.
- Ideal Setup: Clone your new repository, ideally at the root of your project workspace.
-
Clone the Official APM Repository (Recommended for Direct Use & Updates):
- Clone the main APM GitHub Repository directly into your project workspace, ideally at the root.
- This gives you direct access to the latest version and all assets.
-
Manual Copy-Pasting (Basic Usage):
- You can manually copy and paste prompt content from the official APM GitHub repository into your AI assistant as needed. While this approach works, it involves more manual labor and does not fully leverage the agentic capabilities of the framework.
Next Steps:
- Begin with the Introduction to familiarize yourself with the APM framework.
- Proceed to the Getting Started Guide or consult the Quick Start Guide to launch your first APM session.
APM v0.4 includes comprehensive documentation covering all aspects of the framework:
| Document | Description |
|---|---|
| Introduction | Overview of APM concepts, goals, and core framework |
| Getting Started | Step-by-step setup and first session guide |
| Agent Types | Different agent roles and specializations |
| Workflow Overview | Complete workflow walkthrough with process diagrams |
| Token Consumption Tips | Cost optimization strategies and model recommendations |
| Modifying APM | Customization of APM assets and advanced features |
| Context & Memory Management (advanced) | How APM handles context and manages memory across agent instances |
| Context & Prompt Engineering (advanced) | Prompt and context engineering techniques used throughout the framework |
These PDF guides provide detailed, visual explanations of APM's processes, including annotated screenshots, best practices, and practical tips to enhance your understanding:
| Guide | Description | Best For |
|---|---|---|
| Quick Start Guide | Step-by-step walkthrough of your first APM session with annotated screenshots | Beginners & visual learners |
| User Guide | In-depth manual covering advanced usage, optimization techniques, and troubleshooting for common issues | Experienced users |
For a complete documentation index with recommended reading order, see the Documentation Hub.
APM is an open-source project, and your contributions are welcome! Whether it's improving prompts, enhancing documentation, suggesting new features, or reporting bugs, please feel free to open an issue or submit a pull request.
Ways to contribute:
- Report bugs or workflow issues
- Suggest features or improvements
- Improve documentation or guides
- Share customizations/adaptations for specific domains or IDEs
Areas particularly seeking contributions:
- JSON Asset Format Testing: APM v0.4 includes an experimental JSON asset format variant designed for better LLM parsing. If you have access to high token budgets or want to experiment with structured asset validation, testing the JSON format and providing feedback would be valuable for future development. See Context & Prompt Engineering documentation and prompts/schemas/README.md for detailed information and contribution guidelines.
- Ad-Hoc Delegation Guides: The framework currently includes Debug and Research delegation guides, but there's opportunity to create specialized guides for other context-intensive tasks such as testing automation, security analysis, data extraction and more. See prompts/ad-hoc/README.md for contribution guidelines and template patterns.
Please read CONTRIBUTING.md for details on the code of conduct and contribution process.
This project is licensed under the Mozilla Public License 2.0 - see the LICENSE file for full details.
As APM has matured from an experimental framework into a comprehensive multi-agent coordination system with growing commercial interest, its license has been upgraded from MIT to Mozilla Public License 2.0 (MPL-2.0). This change helps protect the Open Source Software community while maintaining full commercial compatibility.
What this means: APM remains completely free and Open Source for all uses (personal, commercial, enterprise). You can build proprietary products using APM, integrate it into commercial IDEs, and create paid services around it. The only requirements are that improvements to core APM files must be shared back with the community, and that you attribute the creators and the APM project as required by the MPL 2.0 license. Make sure to read the LICENSE file for full details.
