Back to directory

Khazix Skills

A personal collection of Agent Skills, structured instruction sets for AI coding agents, covering goal definition, disk cleanup analysis, AI news lookup, project documentation auditing, research report generation and writing in the author's own voice.

Khazix Skills

Khazix Skills is a set of Agent Skills that its author uses daily, following the open Agent Skills standard so any compatible agent, including Claude Code, Codex, Cursor and more than 40 other tools, can load them as structured instruction sets.

The collection includes leader, which turns a vague idea into a concrete goal brief an agent can run against for hours unsupervised; storage-analyzer, which scans a Mac or Windows disk and ranks files for cleanup; aihot, which fetches a daily AI news digest without an API key; neat-freak, which realigns project documentation, CLAUDE.md files and agent memory and audits whether rules were actually followed; hv-analysis, which produces a long research report on a product, company or concept; and khazix-writer, which writes long-form articles in the author's own tone.

Skills install by cloning a named subdirectory of the repository, or by copying a skill's SKILL.md file into an agent that does not support the Agent Skills standard directly.

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.