divergent-thinking

ADHD — a skill for coding agents. Tree-of-thought with pruning, built on the Claude & Codex Agent SDK. Fans out parallel divergent thoughts under different cognitive frames, scores, prunes traps, deepens the survivors. The no-brainer skill for creative and interdisciplinary work.

[𝐍𝐚𝐭𝐮𝐫𝐞 𝐂𝐨𝐦𝐦𝐮𝐧𝐢𝐜𝐚𝐭𝐢𝐨𝐧𝐬] 🤖💡 LiveIdeaBench: Evaluating LLMs' Scientific Creativity and Idea Generation with Minimal Context

Characterising the Creative Process in Humans and Large Language Models

Your AI converges by default. This makes it diverge. Six Claude Code skills for AI-assisted divergent thinking, built on neuroscience.

Turn LLM coding agents (Claude Code, Codex) from next-token predictors into divergent thinkers. A research-grounded cognitive engine + drop-in Claude Code skills for creativity, reasoning & robustness, with 6 honest reproducible benchmarks and 137 verified papers.

A popular Neuro-Symbolic framework combining Causal Inference, Game Theory, and Bayesian Networks to model historical outcomes and their working principles..

An Orchestration Layer for AI to Ideate and Solve Problems With Divergent Thinking Systems

Tree-of-thought brainstorming engine built on the Claude Agent SDK. Fans out parallel ideas under different cognitive frames, scores them, prunes traps, and deepens the survivors.

The creative psychometric item generator: a framework for item generation and validation using large language models

layton-eval is an AI eval benchmark for divergent, out-of-the-box and non-linear reasoning

A hub of reusable agent skills: CrossThink (divergent cross-domain brainstorming methodology), and more. Built for OpenCode, Claude Code, and Codex.

A Claude Code plugin that generates divergent solutions, attacks them, and merges only what passes the tests.

Meta-Analysis: AI's Differential Effects on Creativity Subfactors (Fluency, Flexibility, Originality, Elaboration) - Mode Collapse Hypothesis

Open Agent Skill for Codex and other harnesses to reopen prematurely converged creative work and explore structural alternatives.

发散性思维测试系统 (DAT) — 基于词向量语义距离客观评估创造力，支持多种 Embedding 模型，FastAPI + CLI + 纯静态前端

ADHD preprint — Parallel Divergent Ideation for Coding Agents

DAT-RU: Russian Adaptation of the Divergent Association Task — 21K submissions, psychometrics, practice effects, lexical predictors

三人行 · divergent 3-seat AI discussion. Provider-agnostic (any OpenAI-compatible model). MCP + web + CLI. Tao Lab, Yunnan University.

Do LLMs get more creative as they get better? 21 models, 1800 runs of the Divergent Association Task — and why the leaderboard is mostly an artifact of the scoring model.

A Claude Skill that tells you what to do first about a problem: five solutions from five lens families plus an Occam's Razor anchor, ranked into one first move. Every version is tested against plain Claude, and every round is published, including the losses.