Review user stories with AI
Give a product draft a fresh review for story size, testable acceptance criteria and consistent terminology.
Review user stories with AI before the planning meeting to find questions a polished draft still leaves unanswered. The po-review skill reads the latest product requirements or stories in a fresh context, focusing on sprint-sized scope, testable acceptance criteria and consistent domain language.
It returns suggestions for the weakest passages and a rough clarity score. That score reflects model judgment rather than executed tests, proven feasibility or a readiness certificate. Your team still decides whether the proposed changes fit its actual capacity and technical constraints.
What it does
- Find acceptance criteria without a testable outcome
- Question whether a story fits one sprint
- Spot inconsistent domain terminology
- Surface ambiguity before planning
- Get focused changes for weak passages
- Review a draft outside its original conversation
Install and use
- Load the selected instruction and licence in a compatible agent. Supply the latest PRD.md or STORIES.md in a fresh review context, with relevant capacity and constraints.
- Ask: “Review this delivery-address change story. Identify ambiguous acceptance criteria and inconsistent terms.”
- Read the scope, testability and terminology findings. Keep assumptions separate from known facts when context is missing.
- Discuss the proposed changes with the team and reread the revised draft. A higher score alone does not replace technical review.
FAQ
Source
Made by Mike Consemulder. Original skill: po-review, published on GitHub under the MIT licence.
View on GitHub