writing-documentation-for-llms

Teaches writing effective LLM documentation using competence assumption, progressive disclosure, and feedback loops.

• Guidance for writing clear, actionable docs for LLMs
• Assumes LLMs are smart; only adds what they don’t know
• Uses examples over explanations, specific triggers over vague terms
• Builds feedback loops into workflows for quality control

• Assume competence: challenge every piece of info needed
• Use progressive disclosure: organize like a table of contents
• Prioritize examples over explanations
• Match specificity to task fragility: narrow for critical tasks
• Include TOC in long files for full scope visibility
• Use consistent terminology across files
• Add feedback loops for quality-critical tasks
• Avoid time-sensitive, vague, or nested references

1. 1.Start with core principles: assume competence, match specificity
2. 2.Structure docs with main file and one-level references
3. 3.Use clear descriptions with what + when
4. 4.Include concrete examples before abstract explanations
5. 5.Break workflows into steps with clear validation steps
6. 6.Test across models and iterate based on performance