Extract Cognitive Frameworks From Any Person

Distills cognitive operating systems from public figures: their mental models, decision heuristics, expression DNA. Into executable perspective skills that help you solve problems through another person's thinking framework. Instead of copying advice, you internalize how someone like Charlie Munger or Richard Feynman actually thinks, and apply that framework to your own questions.

1. Describe what you need to distill (a person or topic). 2. The framework runs 6 parallel research agents across writings, conversations, expression patterns, external views, decisions, and timeline. 3. It distills 3-7 mental models plus decision heuristics through triple verification (cross-domain repetition, generative power, exclusivity). 4. The result is a runnable SKILL.md deployed to your AI agent runtime. Output: a cognitive perspective you can activate on demand across 50+ compatible runtimes (Claude Code, Cursor, Codex, OpenClaw, Hermes, etc.).

Access frameworks from the world's strongest thinkers through their actual cognitive structure, not quoted advice, in any runtime you use.

Requires Node 18+ for CLI installation (npx skills add). Skills are standard Agent Skills protocol compliant, compatible with 50+ runtimes. The framework uses markdown plus YAML frontmatter for persistence. Phase 2 uses Python 3.7+ utility scripts for research aggregation and quality checking if running offline validation. Expects WebSearch or similar agent capability for Phase 1 multi-source research.

This distills expressed thought patterns, not hidden intuition or real-time face-to-face instincts. Information is a snapshot through the research date. Public statements may diverge from private positions. The framework works best for people with substantial public records; minimal-footprint figures yield shallow models.