← ClaudeAtlas

pm-refinelisted

Use when refining user stories, writing feature specifications, sizing work (T-shirt), creating subtasks, planning spikes, or generating QA test plans - covers the product management side of development before implementation begins
badrusiddique/enggenie-skill · ★ 0 · Testing & QA · score 70
Install: claude install-skill badrusiddique/enggenie-skill
# enggenie:pm-refine Turn vague feature requests into deployable specifications. This skill takes "I want to build X" and produces a complete spec with acceptance criteria, estimation, subtask breakdown, QA plan, and Jira tickets. It also refines existing stories, plans spikes, and re-estimates work with transparent reasoning. No spec ships with TBD sections. No story gets estimated without showing the math. No feature starts implementation without edge cases documented. --- ## Announcement When this skill is invoked, announce: > I'm using enggenie:pm-refine to generate a feature specification. --- ## Modes This skill operates in four modes. The user's intent determines which mode activates. ### Spec Mode Trigger: "I want to build X", "Write a spec for X", "New feature: X" Generates a complete feature specification from scratch. This is the primary mode and follows the full workflow described below. ### Refine Mode Trigger: "Refine PROJ-1234", "Polish this story", "Tighten the AC on X" Takes an existing story or set of acceptance criteria and improves them. Pulls the current state from Jira (if available), identifies gaps, strengthens acceptance criteria, adds missing edge cases, and updates the ticket. Does not generate a new spec from scratch -- works with what exists. ### Spike Mode Trigger: "Spike: can we use X for Y?", "Research ticket for X", "Time-boxed investigation" Generates a time-boxed research ticket. Every spike has: - A clear question to answe