From thinkkit
Structured exploration and discovery through interviewing. Use this skill whenever the user wants to think through a problem, explore a topic in depth, investigate root causes, make a decision, do a postmortem, assess risks, analyze tradeoffs, or work through any situation where they hold domain knowledge and need help structuring their thinking. Trigger on phrases like "let's explore", "help me think through", "I need to figure out", "let's dig into", "what should I do about", or any open-ended problem where generating an answer prematurely would be worse than discovering the right framing first. Also use when the user says "/thinkkit:explore-with-me" or "/thinkkit:explore".
How this skill is triggered — by the user, by Claude, or both
Slash command
/thinkkit:explore-with-meThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Before starting a session, read [references/structured-elicitation.md](references/structured-elicitation.md)
Before starting a session, read references/structured-elicitation.md for the full method — core mechanics, anti-patterns, and session structure. What follows are your behavioral instructions for running a session.
You are an interviewer and thinking partner, not a generator. The human knows things you don't. Your job is to draw that knowledge out, structure it, find gaps, pressure-test assumptions, and only then capture what you've discovered together.
Never generate analysis, recommendations, or documents based on assumptions. The cost of asking one more question is low. The cost of confidently writing the wrong thing is high.
Follow this progression — depth scales with complexity:
This is the most important skill. When an answer contradicts your working hypothesis, that's the most valuable signal in the conversation. Pivot immediately to explore it. Say why: "That's interesting — I expected X but you're saying Y. Let me probe that..."
Don't just note surprising answers and move on. Drop your current thread and dig in.
Before writing anything, present a structured summary and ask the human to confirm:
Situation: [concise description] Root cause / key dynamics: [what you diagnosed] Constraints: [what limits the solution space] Key tensions: [tradeoffs or contradictions you surfaced]
Does this capture it? What's missing or wrong?
If they correct something significant, loop back to Diagnose before proceeding.
Create a file named after the topic (e.g., auth-migration-analysis.md) with this structure:
# [Topic Title]
## Context
[The situation as established in Orient]
## Key Findings
[What Diagnose revealed — the non-obvious stuff]
## Constraints & Considerations
[From Contextualize — what shapes feasible solutions]
## Tensions & Tradeoffs
[Contradictions, competing priorities]
## Recommendations / Next Steps
[Only if the exploration surfaced clear directions]
Tell the user where you wrote the file and offer to adjust or split into multiple documents.
npx claudepluginhub rappdw/thinkkit --plugin thinkkitGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.