deep-interview - Clarify LLM/VLM research ideas through structured interviews
Uses Socratic questions and ambiguity scoring to turn an unclear LLM/VLM research idea into an executable experiment specification.
Tags
Updated: 2026-10-06Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Parse research ideas
- Explore existing codebases
- Ask targeted questions
- Score ambiguity dimensions
- Track interview state
- Generate experiment specifications
- Bridge to execution skills
Inputs
- Research idea or vague description
- Interview mode options
- Existing source code
- Runtime ambiguity threshold
Outputs
- Interview questions
- Ambiguity score reports
- Saved interview state
- Deep interview specification file
- Execution skill selection prompt
Requirements
- LLM/VLM-capable model
- Haiku model support
- Opus model support
- AskUserQuestion tool
- Task tool
- state_read and state_write tools
- Write tool
- Skill tool