nuwa-skill: A DSH plugin for Distilling How People Think in DeepSeek Harness
alchaincyf/nuwa-skill
Nuwa is a DSH plugin that distills how anyone thinks — mental models, decision heuristics and expression DNA — into ready-to-use perspective skills from just a name.
nuwa-skill is a DSH plugin that turns a notable person's way of thinking into something you can actually call. It targets a common gap: you want a great thinker's judgment framework but only get scattered quotes. Give it a name and it handles research, extraction, and validation on its own, producing a persona Skill covering expression DNA, mental models, decision heuristics, anti-patterns, and honest limits, then writes it into the project skills directory for DeepSeek Harness to load automatically. The output follows the Agent Skills protocol and also works in Claude Code, Codex, Cursor, and other compatible runtimes.
How to Install
dsh plugin --profile <your-profile> add github:alchaincyf/nuwa-skillReplace --profile with the profile name you actually run (see ls ~/.dsh/profiles/): usually web in the browser UI, desktop or desktop-NNN in the desktop app.
- Category
- Skills & Agents
- Platform
- DSH Plugin
- Author
- alchaincyf
- Distribution
- Plugin
nuwa-skill Key Features
nuwa-skill Repo Summary
What Does It Do?
Nuwa (nuwa-skill) is a DSH plugin for DeepSeek Harness that distills how any person thinks into a reusable Skill. It solves the problem of wanting a great thinker's cognitive framework but only getting quote recitations: you give it a name, and it runs research, extraction, and validation automatically to produce mental models and decision heuristics rather than role-play. Built on the open Agent Skills protocol and maintained by alchaincyf under the MIT license, the plugin writes distilled persona Skills into your project's skills directory, where DSH loads them automatically.
Core Features
- One-name distillation: give it a single name and the plugin completes research, extraction, and validation to produce a standalone persona Skill.
- Five-layer extraction: it captures expression DNA (tone, rhythm, word choice), mental models, decision heuristics, anti-patterns and value boundaries, plus honest limits.
- Perspective-based analysis: once distilled, you can ask it to analyze a decision through Munger's lens, grounding answers in cognitive frameworks instead of famous quotes.
- Declared honesty boundaries: every Skill states what it cannot do, such as distilling intuition, capturing sudden change, or reading private thoughts beyond public information.
- Cross-runtime output: generated Skills follow the Agent Skills protocol and also run in Claude Code, Codex, Cursor, and other compatible runtimes.
How to Use This Plugin?
After enabling it in DSH, tell the agent something like "distill Paul Graham" or "build a Zhang Xiaolong perspective Skill", and the plugin starts researching and generating the persona Skill. Once created, invoke it by name, for example "How would Feynman explain quantum computing?" or "Switch to Naval, I'm torn between three things". For a quick one-off, you can also paste the Skill's markdown document straight into the conversation, since it is essentially markdown with YAML frontmatter.
How to Troubleshoot This Plugin?
If a distilled Skill does not show up in DSH, first check whether the persona Skill was written into the project's .dsh/skills/ directory, since DSH only auto-loads content from there. If the agent cannot recognize an installed Skill, verify that your runtime supports automatic Agent Skills loading; when it does not, follow the official guidance and paste the Skill document into the conversation instead.
This page is an independent rewrite of the plugin's official README — for authoritative documentation and the latest changes, refer to the source: alchaincyf/nuwa-skill. The plugin is third-party code that runs on your machine once installed; inclusion does not imply endorsement — please review the source before installing.
DSH Plugin FAQ
