This repository provides a registry of reusable AI agent skills designed to automate repeatable workflows for Chinese knowledge workers. It transforms complex tasks like weekly reviews, note-taking, content analysis, and publishing into machine-readable skills for LLM agents such as Claude Code, Codex, and Gemini, enabling AI to handle structured processes while humans retain judgment and focus on higher-level goals.

How It Works

The system operates as a GitHub-backed registry where individual skills are defined in skills/*/SKILL.md files. A static website generator processes these files, extracting metadata to produce LLM-readable manifests (agent.json), a searchable skill index (skills.json), installation commands, version labels, and update timestamps. This approach makes workflows inspectable by humans and executable by AI agents, facilitating structured knowledge management and automation.

Quick Start & Requirements

To install a skill locally, copy the relevant skill directory (e.g., skills/fire-analysis-card) into your agent's skill directory (e.g., ~/.codex/skills/). Building the website requires npm. The system is designed for integration with "mainstream LLM agents."

• Primary Install Command: cp -R skills/<skill-name> ~/.codex/skills/
• Prerequisites: python3, npm, git.
• Links:
  • Live Site: https://www.twhsi.com/
  • Skill Index: https://www.twhsi.com/skills.json

Highlighted Details

• iMandalArt 2.01: Compresses source material into a CJK-friendly 3x3 Mandala index card with a strict 11-line text contract, optimized for chat interfaces and knowledge capture.
• FIRE Semantic Analysis: Implements a card-box thinking system (Full-D, Index, Route, Evolution) for structured writing and retrieval.
• Keyword Graph View: Extracts eight context-sensitive keywords from text, generating a weighted network graph with definitions and evidence.
• Eight Page Booklet 2.0: Compiles diverse content types into a printable, foldable A4 eight-page booklet, complete with validation artifacts for structure, typography, and layout.

Maintenance & Community

Maintenance involves updating skill files, validating changes, rebuilding the website, and pushing to the main branch to ensure the live site reflects the latest versions. The README does not specify community channels (e.g., Discord, Slack) or notable contributors/sponsorships.

Licensing & Compatibility

The repository's license is not explicitly stated in the README. This omission requires clarification for assessing commercial use or closed-source integration compatibility. The skills are designed for mainstream LLM agents and focus heavily on Chinese (CJK) workflows.

Limitations & Caveats

The project's primary focus is on Chinese knowledge workers and workflows, although some skills support mixed or English text. An English-native version of iMandalArt is planned but not yet implemented. The absence of a specified license is a significant caveat for adoption.