# Analogy Engine

> Analogy Engine

- Skill: `sethdford/analogy-engine` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add sethdford/analogy-engine`
- Raw SKILL.md: https://api.skillmd.com/api/skills/sethdford/analogy-engine/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/analogy-engine

---

# Analogy Engine

Transfer patterns across domains by mapping **structure**, not surface similarity: identify roles, flows, feedback loops, and bottlenecks in a source world, then find the same skeleton in the target problem.

## When to Use
- Domain experts have exhausted conventional playbooks and options feel interchangeable.
- You want inspiration from biology, logistics, hospitality, games, or other far fields.
- A tempting metaphor appears—use this skill to **verify** it before betting on it.

## Workflow
1. State the problem in abstract terms (inputs, actors, state, goals, failure modes).
2. Pick a distant domain (biomimicry, another industry, or “What would X do?”—e.g., a restaurant handling a software reliability issue).
3. Map **structural** correspondences: what plays which role? Where is the queue, the filter, the catalyst?
4. Translate one or two mappings into concrete tactics; list where the analogy **breaks** and adjust.
5. Prefer analogies backed by shared dynamics, not shared vocabulary.

## Examples
**Example 1:** Latency spikes: kingfisher beak → smooth boundary transitions; test whether “gradual impedance change” applies to API batching vs sharp bursts.

**Example 2:** Onboarding drop-off: compare to restaurant seating—reservation, waitlist, and “first drink” map to pre-auth, staggered rollout, and instant value.