Context Engineering Advisor

If you've ever dumped an entire PRD into Claude and gotten mediocre output, this walks you through why. It distinguishes context stuffing (jamming everything in) from context engineering (treating AI attention as scarce). The diagnostic centers on five questions like "what specific decision does this support?" and introduces the Research→Plan→Reset→Implement cycle to prevent context rot. Honest take: this is less about prompts and more about information architecture. If you're burning tokens without seeing better results, or your AI outputs feel inconsistent, the structured diagnosis here will likely surface whether you're hoarding context or actually engineering it.

Install to Claude Code

npx -y skills add deanpeters/product-manager-skills --skill context-engineering-advisor --agent claude-code