xskill Wiki · docs 🥥

Overview

Quickstart

$ pip install xskill          # Python 3.9+
$ xskill serve                # writes ~/.xskill/config.yaml, then exits

skill_dir: ~/.xskill/skill

llm:
  base_url: https://api.deepseek.com
  model:    deepseek-v4-flash
  api_key:  YOUR_KEY

embedding:
  base_url: https://api.deepseek.com
  model:    deepseek-embedding
  api_key:  YOUR_KEY
  dim:      0

$ xskill registry add /path/to/trajectories

Concepts

Five terms cover the whole system:

Term	Meaning
Trajectory
Atom
Skill
Canary
UX score

How it works

Architecture

Supported agents

Trajectory ingest and Skill install are implemented per ecosystem:

Agent	Status	Trajectory ingest	Skill install
Claude Code	verified	`~/.claude/projects/`	`~/.claude/skills/<name>/`
Codex CLI	verified	`~/.codex/sessions/`	`~/.agents/skills/<name>/`
OpenCode	verified	`~/.local/share/opencode/opencode.db`	`~/.agents/skills/<name>/`
OpenClaw	not well tested	`~/.openclaw/agents/`	`~/.agents/skills/<name>/`
Cursor	not well tested	`~/.cursor/projects/*/agent-transcripts/`	`~/.cursor/skills/<name>/`
Any other agent	manual	`xskill.adapters.submit_trajectory`	copy / symlink `SKILL.md`

Team mode

$ xskill serve --server                        # prints a join token
$ xskill connect <host:port> --token <token>

—
—
—
—
—

CLI reference

xskill serve [--host 0.0.0.0] [--port 8000] [--server]

xskill registry <add|remove|list> [path] [--label LABEL]

xskill search <traj|skill> <query> [-k/--top-k N]

xskill connect [host:port] [--token TOKEN] [--label LABEL]

Configuration

llm:
  base_url: https://api.deepseek.com
  model:    deepseek-v4-flash
  api_key:  YOUR_KEY
  rate_limit:
    rpm: 60          # match your provider plan
    tpm: 100000      # optional
    burst: 10        # optional; default = ceil(rate/6)

xskill documentation