From workflow-toolkit
This skill should be used when the user asks to "grill me", "stress-test my plan", "challenge my design", "poke holes in this", "review my architecture", "play devil's advocate", "find the flaws", "what am I missing", or wants critical feedback on a plan, design, or idea before implementation.
How this skill is triggered — by the user, by Claude, or both
Slash command
/workflow-toolkit:grill-me*.md*.txtThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Interview the user relentlessly about every aspect of their plan until reaching shared understanding. Break the plan into categories and resolve dependencies between decisions one by one.
Interview the user relentlessly about every aspect of their plan until reaching shared understanding. Break the plan into categories and resolve dependencies between decisions one by one.
Ask the user to describe their plan, design, or idea in detail. If the plan is already in the conversation context, summarize the understanding and confirm before proceeding.
Break the plan into categories. Common categories include:
Adapt categories to the plan. Not every plan is technical — the same rigor applies to product plans, process changes, or organizational decisions.
For each category:
Keep the tone constructive but relentless. The goal is to surface blind spots, not to criticize. Challenge assumptions directly. If something sounds hand-wavy, say so.
After all branches are resolved, produce a structured decision log:
## Decision Log
### Category: [Category Name]
| # | Question | Decision | Rationale |
|---|----------|----------|-----------|
| 1 | [Question asked] | [Answer agreed upon] | [Why this was chosen] |
| 2 | ... | ... | ... |
### Category: [Next Category]
| # | Question | Decision | Rationale |
|---|----------|----------|-----------|
| 1 | ... | ... | ... |
### Open Questions
- [Any unresolved items that need further investigation]
### Key Risks
- [Risks identified during the interrogation that the user should monitor]
Present the decision log to the user. Ask if anything was missed or if any decisions need revisiting. The interrogation is complete only when the user confirms the log is accurate.
Once the decision log is confirmed, suggest: "When ready to formalize this into a requirements document, ask me to write a PRD."
npx claudepluginhub pau-vega/ai-devkitGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.