Day 7: Prompt Engineering (TBU)

Read and apply this to my CLAUDE.md: https://www.humanlayer.dev/blog/writing-a-good-claude-md

⏺ Fetch(https://www.humanlayer.dev/blog/writing-a-good-claude-md)
  ⎿  Received 66.4KB (200 OK)

⏺ Now let me read the current CLAUDE.md and analyze it:

⏺ Read(CLAUDE.md)
  ⎿  Read 181 lines

⏺ ★ Insight ─────────────────────────────────────
  Key Takeaways from the HumanLayer Blog Post:
  1. LLMs are stateless - CLAUDE.md is the primary onboarding mechanism
  2. Less is more - Keep under 300 lines (ideally ~60 lines); Claude reliably follows 150-200 instructions max
  3. Three-part framework: WHAT (structure), WHY (purpose), HOW (workflows)
  4. Progressive disclosure - Reference docs rather than embed everything
  5. Don't use Claude as a linter - Remove style instructions; use deterministic tools
  6. Only universally applicable guidance - Avoid task-specific content
  ─────────────────────────────────────────────────

  Let me analyze your current CLAUDE.md (181 lines) against these principles: