# Pattern Recognition

> Pattern Recognition

- Skill: `sethdford/pattern-recognition` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add sethdford/pattern-recognition`
- Raw SKILL.md: https://api.skillmd.com/api/skills/sethdford/pattern-recognition/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: sethdford (https://skillmd.com/u/sethdford)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/sethdford/pattern-recognition

---

# Pattern Recognition

Map the **structure** of a problem to structures you have seen elsewhere, then borrow *mechanisms* (not labels). The win is transfer; the failure mode is superficial analogy.

## When to Use
- Novel situations that “rhyme” with past domains; systems with loops, incentives, networks, or coordination; when the user wants reusable intuition.

## Workflow
Ask: **“Where have I seen this structure before?”** Scan for: feedback loops; network effects; diminishing returns; threshold/step-function behavior; tragedy of the commons; coordination games; principal–agent misalignment.

For each candidate pattern, state **the matching mechanism** (who acts, what reinforces what, what is conserved/scarce) and **a falsifiable check**—what observation would disprove the analogy. Reject matches based on surface vocabulary alone.

Prefer **one strong structural match** plus implications over a laundry list of buzzwords.

## Examples
**Example 1:** Org silos hoarding data → tragedy of the commons / local incentives: private upside, shared downside; fix needs metrics and governance, not “more culture slides.”

**Example 2:** Latency SLO debates → diminishing returns + threshold effects: past a knee, small wins don’t change UX; verify with percentiles and user journeys, not mean latency alone.