Research
nuwa-skill avatar

nuwa-skill

Nuwa (女娲): Auto-research any person to extract their core mental models, decision heuristics, and expression DNA to generate a runnable perspective skill.

Introduction

Nuwa is an advanced cognitive engineering engine designed to capture the essence of a person's thinking process. Rather than simply summarizing biographies, Nuwa acts as a multi-agent swarm that decomposes a target's intellectual output into structured mental models, decision heuristics, and stylistic 'DNA'. It is the perfect tool for researchers, strategists, and knowledge workers who need to stress-test their ideas against different professional perspectives or simulate high-level advisory dialogues.

  • Multi-Agent Parallel Research: Simultaneously analyzes writings, long-form conversations, decision records, and external criticism using six specialized agents.

  • Cognitive Framework Extraction: Identifies core belief systems, decision-making biases, and recurrent analogies to build a runnable 'cognitive operating system.'

  • Expression Style Analysis: Documents unique language patterns, recurring vocabulary, and humor style to ensure high-fidelity role-play and simulation.

  • Self-Contained Skill Generation: Produces a structured directory containing deep-dive research files (01-writings.md, 02-conversations.md, etc.) and a final SKILL.md deliverable.

  • Honest Constraint Mapping: Explicitly identifies what a skill cannot do and preserves information contradictions rather than smoothing them over.

  • Input Requirements: Provide any name or topic; the system prompts for focus direction, purpose (advisor, role-play, etc.), and updates.

  • Research Integrity: Prioritizes primary sources (books, interviews, transcripts) over secondary accounts, with a strict source blocklist to ensure high-quality, authoritative data ingestion.

  • Practical Usage: Ideal for creating an 'AI mentor' or a 'sparring partner' that mimics specific thought leaders, authors, or historical figures.

  • Constraints: The skill performance depends on source availability; it provides quality summaries even with limited data but flags information gaps explicitly to prevent hallucination.

Repository Stats

Stars
4,410
Forks
1,199
Open Issues
7
Language
Python
Default Branch
main
Sync Status
Idle
Last Synced
Apr 28, 2026, 11:05 AM
View on GitHub