oh-my-ai/discovery-interview
Interactive discovery interview → structured product spec. Triggers: spec, PRD, requirements, scoping, brainstorming, new project. Uses AskUserQuestion; WebSearch/WebFetch when user wants research. Outputs user stories, acceptance criteria, constraints.
99
100%
Does it follow best practices?
Impact
96%
2.28xAverage score across 3 eval scenarios
Advisory
Suggest reviewing before use
Evaluation results
97%
64%Discovery Interview Simulation: Fitness App
Interview depth and uncertainty options
Minimum question count
100%
100%
Uncertainty option present
66%
80%
Research loop included
0%
100%
All 7 phases present
30%
100%
Phase 7 handoff options
0%
100%
Spec file date-prefixed name
0%
100%
Spec has required sections
33%
100%
Completeness check before spec
0%
100%
94%
39%Discovery Interview: Real-Time Collaboration Tool
Conflict resolution and category coverage
Conflict surfaced explicitly
100%
100%
Conflict resolution options
100%
100%
Multiple categories covered
100%
100%
Minimum question count
100%
100%
Uncertainty option in every question
0%
66%
Completeness check gate
0%
100%
No silent TBDs in spec
100%
90%
Spec file naming convention
0%
100%
Spec uses template structure
0%
100%
99%
61%Discovery Interview: Internal Developer Portal
Spec template structure and handoff phase
Executive Summary present
40%
100%
Problem Statement present
100%
100%
Success Criteria present
100%
100%
User Personas present
100%
100%
User Journey present
0%
100%
P0/P1/P2 requirements
0%
100%
Technical Architecture present
100%
100%
Out of Scope section
100%
100%
Appendix with research
0%
100%
Handoff question in log
0%
100%
Summary confirmed before spec
0%
100%
Completeness check documented
0%
100%
Spec docs/specs/ path
100%
100%
Uncertainty option in all questions
20%
80%
- Evaluated
- Agent
- Claude
- Model
- Claude Opus 4.6