Skill

Evolve Ideas Through Structured Competition

A round-based idea-evolution system that scores, cross-pollinates, and mutates competing ideas through structured rounds.

Works with github

91
Spark score
out of 100
Updated 3 months ago
Version 13.1.0
Models
claude

Add to Favorites

Why it matters

Systematically evaluate and develop scattered ideas by treating them as competing organisms. This skill scores, selects, crosses, and evolves concepts through structured rounds to surface the strongest, most viable ideas.

Outcomes

What it gets done

01

Structure and score raw ideas based on novelty, feasibility, value, logic, cross-potential, and verifiability.

02

Facilitate cross-pollination of ideas to generate unexpected hybrid concepts.

03

Incorporate external stimuli to trigger mutations and spawn new idea variations.

04

Manage the lifecycle of ideas from seed to validated or dormant states.

Install

Add it to your toolbox

Run in your project directory:

curl -fsSL https://spark.entire.vc/get/ag-idea-darwin | bash

Overview

Idea Darwin Engine

A round-based idea-evolution system that scores ideas across six dimensions, cross-pollinates them into hybrids, and mutates them via external stimuli through a structured lifecycle. Use it when managing many scattered ideas needing systematic evaluation or structured iteration rather than one-shot brainstorming.

What it does

A round-based idea-iteration system that treats ideas as competing organisms on an "evolution island" rather than filing them away to rot: each round, the fittest ideas get deepened through structured research - filling logical gaps, clarifying paths, identifying risks - different ideas cross-pollinate into unexpected hybrids, and external stimuli, such as industry news, theories, or conversations, trigger mutations into entirely new ideas. Every idea gets a structured "species card" tracking its core question, full description, parent and child lineage, and a change history, scored across six weighted dimensions - feasibility, value, logic, and verifiability at 20% each, novelty and cross-potential at 10% each - and moves through a lifecycle from seed to exploring, refining, crossing, validated, or dormant, with the user always holding final say over life-or-death decisions.

/idea-darwin round

Usage starts by writing ideas into an ideas.md file, initializing the island (/idea-darwin init --budget 8 --actions 3), then running evolution rounds (/idea-darwin round 3) and managing individual ideas (/idea-darwin dormant IDEA-0005, /idea-darwin wake IDEA-0005) - continuing to feed the island with new ideas and environmental stimuli in a separate stimuli.md file over time.

When to use - and when NOT to

Use it when you have many scattered ideas needing systematic evaluation, want to discover unexpected connections between ideas from different domains, need structured iteration rather than one-shot brainstorming, or want a scoring framework to prioritize investment.

Inputs and outputs

Input: raw, unstructured ideas written into ideas.md, plus optional external stimuli in stimuli.md. Output: structured species cards with six-dimension scores, cross-pollinated idea hybrids, and lifecycle-stage transitions, driven by round-based commands.

Integrations

Available as an open-source GitHub repository, installable via ClawHub (clawhub install idea-darwin), and supports English, Chinese, and Japanese.

Who it's for

Anyone managing a large, scattered set of ideas who wants structured, competitive iteration - scoring, cross-pollination, and mutation - rather than a static idea-storage tool.

Its best practices cut both ways: write ideas as rough as you want since the system structures them, and keep adding external stimuli to prevent ideas from converging on the same handful of directions - but don't over-curate the initial idea list, since letting evolution do the filtering is the point, and don't skip the "Decisions Needed" section of a briefing just because the system only recommends rather than decides on the user's behalf. The six-dimension score is weighted rather than averaged flatly: feasibility, value, logic, and verifiability are each worth twice as much as novelty or cross-potential in the overall total.

FAQ

Common questions

Discussion

Questions & comments · 0

Sign In Sign in to leave a comment.