Back to directory

Cangjie Skill

A tool that distills methodologies from books, long-form videos and podcasts into callable AI agent skills, with a companion website for browsing and installing them.

Cangjie Skill

Cangjie Skill extracts methodologies from long-form source material such as books, videos and podcasts and compiles them into installable AI agent skills. Extraction first produces a stable, content-addressed capability bundle of cards and metadata, then compiles that bundle into either a single router-style skill or a pack of a router plus several standalone skills.

A local toolchain script covers diagnostics, compilation, replanning output, incremental updates, repair, rollback, evaluation and benchmarking, with content-addressed preprocessing, source diffs, impact analysis, transactional patches and snapshots supporting safe changes over time.

A companion website provides visual browsing of published skill packs, a usage guide and a contribution submission form, while this repository remains the source for the project's code, methodology and templates.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.