PM Product Lifecycle Management Assistant

Problem

Product discussions often become fragmented: a rough PRD is drafted, requirements keep shifting, research notes scatter through chat history, and execution lacks clear acceptance criteria. pm-skills structures vague product problems into traceable phases and prevents the workflow from jumping from “discuss” to “generate docs”.

How it works

• Guided lifecycle: it follows Ask → Plan → Execute → Verify → Transition, with BLOCKING gates that require confirmation before moving on.
• Command-driven workflows: core commands such as /pm-init, /pm-plan, /pm-execute, /pm-verify, and /pm-transition load specific workflows and templates; domain commands like /pm-research, /pm-prd, /pm-strategy, and /pm-release handle focused product tasks.
• Templates and state: outputs are stored in files such as PROJECT.md, ROADMAP.md, STATE.md, REQUIREMENTS.md, PRDs, research reports, and release docs to preserve context.
• Capability bridging: deep research can be delegated to deerflow-skill, while knowledge capture can be delegated to llm-wiki-skill; the skill does not reimplement search or knowledge base features.

Boundaries

It fits staged product discovery, PRD writing, strategy, and release preparation, but it is not a one-shot generator for complete product solutions. It works best when users actively confirm requirements, supply business constraints, and accept that templated output may be less exploratory than free-form brainstorming.