Skip to main content

Introspection Output

ShedBoxAI's introspection generates LLM-optimized markdown documentation designed specifically for AI configuration assistance.

Generated Documentation​

The introspection command produces a single comprehensive markdown file containing:

LLM-Optimized Format​

  • Structured for AI consumption: Clear sections and consistent formatting
  • Token estimation: Helps AI understand dataset size constraints
  • Processing recommendations: AI-generated suggestions for optimal operations
  • Sample data: Representative examples for context understanding

Content Sections​

  • Header with metadata: Analysis timestamp and success statistics
  • LLM processing notes: Context window warnings and dataset insights
  • Data source analysis: Detailed schema information and statistics
  • Relationship detection: Automatically identified connections between sources
  • Operation recommendations: Specific suggestions for ShedBoxAI operations

Usage with Claude Code​

The markdown output is specifically designed for use with AI assistants:

# Generate introspection report
shedboxai introspect config.yaml --include-samples -o analysis.md

Then paste the entire analysis.md content to Claude Code with requests like:

  • "Create a ShedBoxAI configuration from this analysis"
  • "Build a pipeline that filters and aggregates this data"
  • "Generate a report configuration based on these sources"

Output Customization​

Limited customization options are available:

Include Sample Data​

shedboxai introspect config.yaml --include-samples

Adds representative data examples to help AI understand data structure.

Custom Output Path​

shedboxai introspect config.yaml -o custom_analysis.md

Saves the report to a custom location.

Next Steps​