Skip to content
CAI
Software that uses CAICheck a score

alchaincyf/nuwa-skill

50.4

Weak · 18 September 2026

1.2k

lines of production code

Python

primary language

1

measurement over time

CAI band scale
CAI lens gauges

What this system is

This system is a persona-based research assistant framework that defines detailed thinking models and behavioral protocols for specific historical figures. It supports these personas with automated utility scripts for processing multimedia content, such as fetching and cleaning subtitles, and for validating the quality and sourcing of the generated research materials.

Features

Add utility scripts for subtitle processing and content quality checks

A new scripts directory has been introduced containing four utility tools: download\_subtitles.sh for fetching YouTube subtitles (prioritizing human-generated over auto-generated, with Chinese and English language preferences), srt\_to\_transcript.py for cleaning SRT/VTT subtitle files into plain text, merge\_research.py for aggregating research findings from multiple agents into a summary table, and quality\_check.py for validating SKILL.md files against specific content standards such as mental model counts, source attribution, and honesty boundaries.

scripts · high confidence

New perspective skills and fidelity scorecards for Andrej Karpathy, Elon Musk, and Richard Feynman

Added three new 'perspective' skills for Andrej Karpathy, Elon Musk, and Richard Feynman, each including a detailed SKILL.md defining their thinking frameworks, behavioral rules, and an 'Agentic Protocol' for research-backed responses. Each skill is accompanied by a FIDELITY.md scorecard (rated A-grade) and extensive research references to ground the persona's views and style.

examples · high confidence

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

How this codebase got here

Baseline

  • First survey — no prior run to compare against. CAI 50.

Lenses

  • Code Health 94
  • Architecture 69
  • Maturity 58
  • Readiness 29
  • Security 100

Changes since last survey

  • 58 commits — 51 feature/other, 7 fixes

By area

  • (root) — 43 commits
  • (repo) — 2 commits
  • assets/banner.svg — 2 commits
  • examples/steve-jobs-perspective — 2 commits
  • examples/sun-yuchen-perspective — 2 commits
  • .github/scripts — 1 commit
  • assets/hero.gif — 1 commit
  • examples/andrej-karpathy-perspective — 1 commit
  • examples/elon-musk-perspective — 1 commit
  • examples/mrbeast-perspective — 1 commit
  • examples/x-mastery-mentor — 1 commit
  • promo/cards — 1 commit

Notable commits

  • fix: fix README: 13位→14位人物,补张雪峰/孙宇晨表格行
  • fix: fix README: 13位→14位人物,补张雪峰/孙宇晨表格行
  • fix: fix README: update examples count from 7 to 13+1, fix '六层' to '五层'
  • fix: fix SKILL.md: add scripts/ to directory structure diagram
  • fix: fix timeline: nuwa predates colleague-skill, not inspired by it
  • fix: fix: add missing research files to x-mastery-mentor example
  • fix: fix: update 小红书 display name from 花生 to 花叔
  • change: Add George Hotz skill to community index (#69)
  • change: FanBox 宣传位更新:主 slogan 改为「Coding Agent 的驾驶舱」+ banner 换 .jpg
  • change: Initial release: Nuwa — distill how anyone thinks
  • change: Merge branch 'auto-optimize/20260701-nuwa'
  • change: README 增加 FanBox 宣传 banner(看得见的 vibe coding)
  • change: add 5 perspective skills + update SKILL.md to latest
  • change: add Zhang Xuefeng example - high public awareness case
  • change: add multilingual READMEs and Star History section
  • change: add promo/ — 宣传素材(deck/hero film/landing/quote cards) by huashu-design
  • change: add scripts/ directory with utility scripts, update SKILL.md paths
  • change: add sun-yuchen-perspective: 孙宇晨.skill — 注意力收割操作系统
  • change: add: hero banner
  • change: add: 达尔文.skill promotion with banner and install command
  • …and 38 more

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

Survey your own repository

alchaincyf/nuwa-skill was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.

About this page

  • The score is its most recent published measurement, taken on 18 September 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
  • Measured at commit fe0374687037c4cc51a65c1e0c145afe2981dc69 — the exact code this score is about.
  • Scored under rubric-2026.09.15 — the same rubric and the same method as every other entry in this index.
  • Measured by watchdog.canine.dev using codehealth-analyzer preprod-5d04157a340d.