This skill runs a devil’s advocate review of a feature idea for a product that already exists. It looks at the idea from the product manager, designer and engineer points of view and lists what could go wrong in four risk areas: value, usability, viability and feasibility.
For each assumption it notes what could fail, how confident the team is (high, medium or low) and a way to test it. The aim is to strengthen the idea before work starts.
What You Get
- A list of risky assumptions grouped by value, usability, viability and feasibility
- A confidence rating of high, medium or low for each assumption
- A suggested way to test each assumption
- Input that feeds directly into assumption prioritization
Common Use Cases
- Stress-testing a feature idea before it enters the roadmap
- Preparing assumptions for an assumption-mapping workshop
- Running a quick risk assessment on a PRD or design the team already has
What Makes It Unique
It is scoped to features in an existing product, so it uses the four classic product risks and a three-perspective devil’s advocate pass. Its sibling skill for new products widens the list to eight categories. The tone is constructive: the goal is to strengthen the idea.
Under The Hood
The skill is a prompt only. It ships no scripts and runs no commands. It tells the agent to read any files the user supplies (designs, PRDs, research) and then reason step by step. It does not call web services, fetch pages or write files of its own.