← Back to Catalog
Agent Skill

thinking-patterns

Longitudinal cognitive pattern analysis across months of recorded conversations.

Longitudinal cognitive pattern analysis across months of recorded conversations. Extracts evidence-based dimensions from Fathom transcripts, synthesizes cross-session patterns, and detects blind spots you can't see yourself.

What it does

Processes months of Fathom meeting transcripts through a multi-stage parallel agent pipeline: extracts your speech, analyzes 12 cognitive dimensions per transcript (Burns’ distortions, hedging, metaphors, decision heuristics, etc.), then synthesizes cross-session patterns including contradictions, blind spots, and execution gaps between stated priorities and actual focus.

Key features

  • 12 extraction dimensions — cognitive distortions, problem framing, conceptual metaphors, hedging/certainty, code-switching, decision moments, emotional indicators, avoidance, agency language, competing commitments, role markers, energy signals
  • 10 output sections — recurring narratives, problem framing, metaphors, decision heuristics, avoided topics, contradictions, energy patterns, role shifts, execution gap, biases
  • Multi-agent architecture — 10-13 parallel Sonnet agents for extraction, 5 for synthesis
  • Evidence-based — every finding includes dated quotes from transcripts
  • Blind spot detection — “The 5 Things You Don’t See” summary

When to use

Quarterly or after accumulating 10+ transcripts, when you want an honest external perspective on your cognitive patterns, decision-making habits, and the gap between what you say you prioritize and what you actually do. Costs approximately $3.50 per run.