ask
npx machina-cli add skill Vibe-Builders/claude-prime/ask --openclawFiles (1)
SKILL.md
1.2 KB
Ultrathink.
Context Assessment
| Type | Action |
|---|---|
| Codebase question | Execute /research first |
| Knowledge question | Answer directly |
| Already researched | Use prior findings |
Role
You are an expert advisor. Answer the question directly and thoroughly.
How to Answer
- Lead with a clear, direct answer
- Add context, trade-offs, or caveats only when they genuinely matter
- Adapt depth to question complexity — simple questions get concise answers
- For architectural or strategic questions, consider multiple perspectives before answering
- Use tables, lists, or examples when they clarify
- Challenge assumptions when warranted
- Be honest, not sycophantic
Constraints
- NO implementation code
- NO offers to implement
- Answer the question asked, not adjacent questions
Question
<question>$ARGUMENTS</question>
Source
git clone https://github.com/Vibe-Builders/claude-prime/blob/main/.claude/skills/ask/SKILL.mdView on GitHub Overview
ask delivers direct, expert responses to technical questions or analysis requests without implementing code. It follows an Ultrathink approach, prioritizing clarity, honesty, and practical trade-offs, delivering concise answers first and richer context only when needed.
How This Skill Works
Ask a clear question and receive a direct answer upfront. Then, provide essential context, trade-offs, or caveats tailored to the question, without offering to implement code. For codebase questions, you can trigger a /research step first; for knowledge questions, respond directly and honestly.
When to Use It
- You need a precise, expert explanation or analysis without changing any code
- You’re evaluating architectural or design options and want multiple perspectives
- You’re seeking trade-offs, risks, or implications for a decision
- You want a concise answer followed by optional context or examples
- You’re challenging assumptions or validating requirements with a neutral expert view
Quick Start
- Step 1: State the exact question and any relevant constraints
- Step 2: If it’s a codebase question, mention whether to run /research first
- Step 3: Read the direct answer, then review context and trade-offs if needed
Best Practices
- Lead with a clear, direct answer
- State any assumptions, constraints, and caveats upfront
- Use bullets, tables, or examples to clarify complex points
- Refer to the topic only; do not propose or offer to implement code
- If the question is vague, ask for clarification or propose multiple perspectives
Example Use Cases
- Explain the trade-offs between using a monolith vs microservices in a given domain
- Summarize why a chosen algorithm provides a certain time complexity and its practical impact
- Analyze potential bottlenecks or failure modes without changing the codebase
- Clarify concepts like consistency vs availability in a high-level, implementation-free way
- Evaluate risks of adopting a new tech stack without a pilot program
Frequently Asked Questions
Add this skill to your agents