Back to directory

CC Harness Skills

Six portable agent skills for memory consolidation, context compression, verification, and multi-agent coordination, distilled from a coding-agent codebase and rewritten to install into Claude Code, Codex, or OpenClaw.

CC Harness Skills

CC Harness Skills packages six installable skills that address problems common to coding-agent setups: keeping memory useful without storing stale code facts, compressing long conversations without losing corrections a user has made, verifying claimed completion instead of trusting it, coordinating multiple agents without polluting the main context, and adding limited proactive behavior safely.

Each skill ships as a host-agnostic bundle with supporting scripts, distilled and rewritten from a publicly mirrored coding-agent codebase so it does not depend on that codebase's private runtime internals. A basic bundle check script validates the skills after installation.

Skills install with a one-line command for any compatible agent, or can be copied individually into Claude Code's, Codex's, or OpenClaw's skill directories, and can be published to a community registry after passing the included checks.

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.