PMs Get Open-Source AI Agent for Frameworks

The project's creator, Paweł Huryn, is a product advisor with over 15 years of experience who runs "The Product Compass," a top technology newsletter on Substack with over 105,000 subscribers. His work focuses on turning established product management theory into actionable, data-driven practices. The "PM Skills Marketplace" is built as a collection of agentic workflows, which execute sets of prompts in sequence rather than operating as fully autonomous agents. This structure is designed to guide product managers through specific, proven methodologies step-by-step, providing analytical frameworks instead of just generic text generation. The encoded frameworks from Marty Cagan are central to his "Product Operating Model," which pushes for empowered, cross-functional teams focused on solving problems and driving outcomes, not just shipping features. Cagan's principles were distilled from his experiences at major tech companies like Netscape and eBay. Teresa Torres's "Opportunity Solution Tree" is another core component, providing a visual framework to map a desired outcome to the user needs and opportunities that could achieve it. This method is designed to structure the messy process of product discovery and ensure teams are solving validated customer problems before building solutions. The system operates through specific slash commands that trigger multi-step workflows. For example, a user can type `/discover` to initiate a chain of four distinct skills: brainstorming ideas, identifying assumptions, prioritizing those assumptions, and then brainstorming experiments for validation. This tool joins a growing ecosystem of AI for product managers, but it differs from general-purpose agents like Auto-GPT or integrated features in tools like Jira and Amplitude. Its specific value lies in being an open-source, domain-specific toolkit trained on the explicit language and logic of product management, including concepts like RICE scoring and OKRs.