agent-ecosystem

Skillxp — AI skill for Claude Code

AI community

Skill invocation runner: install an Agent Skill in a fresh fixture, invoke an agent harness headlessly, and report what actually reached the model, with transcript evidence.

How to install Skillxp

This entry records only its repository, not the path inside it, so there is no exact command to give. Open agent-ecosystem/skillxp and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Skillxp does

Skill invocation runner: install an Agent Skill in a fresh fixture, invoke an agent harness headlessly, and report what actually reached the model, with transcript evidence.

Alternatives in AI

Better Harness — Review this project's AI coding workflow with Better Harness and generate a durable report 2.1k ★
LongHorizon Harness — The long-horizon computer-use harness 1.3k ★
Rb Setup — First-time setup 1.3k ★

README

skillxp

Skill invocation runner: install an [Agent Skill](https://agentskills.io) in a fresh fixture, invoke an agent harness headlessly, and report what actually reached the model, with transcript evidence.

Skill authors publish for 25+ platforms that each load, present, and manage skills differently, and mostly invisibly. skillxp makes that behavior observable: it stages a skill on a real harness and answers "what did the platform actually do with it" from the session transcript, not from the model's self-reporting. It builds on [agentsummons](https://github.com/agent-ecosystem/agentsummons) (headless invocation) and [agentminutes](https://github.com/agent-ecosystem/agentminutes) (transcript parsing), and adds the third layer of lore: how each harness discovers, activates, and records skills.

skillxp renders no verdicts. It produces observations; graders consume them. The first consumer is the [agent-skill-implementation](https://github.com/agent-ecosystem/agent-skill-implementation) loading benchmark, whose checks and verdict logic live in that repo's `benchmark-runner/`.

Status

Early development. Supported harnesses: Antigravity CLI, Claude Code, Codex CLI, and GitHub Copilot CLI. Run `skillxp harnesses` to see the harness versions your installed release was validated against, and `skillxp doctor` to compare them with what you have installed (newer harness releases usually keep working; validation records coverage, not a compatibility bound). Results reflect **headless** behavior, which may differ from interactive use.

Install

brew install agent-ecosystem/tap/skillxp
# or
npm install -g skillxp
# or
pip install skillxp
# or
go install github.com/agent-ecosystem/skillxp/cmd/skillxp@latest

As a Go library: `go get github.com/agent-ecosystem/skillxp`. The harnesses you observe must be installed and authenticated. On Windows, use pip or a [release binary](https://github.com/agent-ecosystem/skillxp/releases); the npm package temporarily ha