Claude Commands - Ask

ask.md

description
Answer questions (/ask)

Answer the user's questions by conducting thorough research and investigation.

IMPORTANT: Never modify files - only provide answers based on research

Execution Conditions

You must verify the following file exists before proceeding:

• ./tmp/context.md - User's questions

If the file is missing:

• Stop the process immediately
• Notify the user which condition failed
• Do not proceed with implementation

Execution Process

Phase 1. Load Questions

• Read the questions from ./tmp/context.md
• Parse the content to identify individual questions
• Each question may be separated by "=====" or similar delimiters
• Extract file references and context (e.g., tmp/plan.md:L144)

Phase 2. Research

• For each question, use sub-agents to conduct thorough investigation
• Use Task tool with general-purpose agent for complex research
• Search codebase for patterns, implementations, and context related to the questions
• Use Web search tool for external documentation and examples
• Use Context7 MCP for library/framework usage
• Use Serena MCP for semantic code search

Phase 3. Investigation & Analysis

• Read referenced files and understand context
• Search for related code patterns and implementations
• Look for documentation and examples
• Prevent hallucination by following these strict guidelines:
  • Only use information directly found in the codebase, documentation, or research results
  • Use direct quotes from source files when making claims about code or documentation
  • If no supporting evidence exists for a claim, explicitly say "I don't know" or "No evidence found"
  • Clearly distinguish between verified findings and any speculation
• Think harder before you answer the questions

Phase 4. Answer The Questions

• IMPORTANT: Apply critical thinking to question assumptions before finalizing answers

Provide answers to user:

• Provide clear, concise answers to each question
• Reference specific files and line numbers when relevant (format: file_path:line_number)
• Reference specific URLs based on the answers
• Include code snippets or examples when helpful
• Separate answers for multiple questions clearly
• Be specific about what was found vs. what is speculation

For multiple questions, format as:

## Question 1: [question summary]
[Answer with specific references and findings]

## Question 2: [question summary]
[Answer with specific references and findings]