Skill

The open agent skills ecosystem — rewritten in Rust.

A drop-in, feature-equivalent Rust replacement for the Vercel skills CLI. Single static binary, zero runtime dependencies, full API compatibility.

Supports Cursor, Claude Code, Windsurf, Codex, and 35 more.

Install a Skill

skills add qntx/skills

Install the CLI

Shell (macOS / Linux):

curl -fsSL https://sh.qntx.fun/skill | sh

PowerShell (Windows):

irm https://sh.qntx.fun/skill/ps | iex

Cargo:

cargo install skills-cli

From Source:

git clone https://github.com/qntx/skill.git
cd skill
cargo install --path skills-cli

Overview

skill manages portable instruction sets ("skills") for AI coding agents. A skill is a directory containing a SKILL.md file with YAML frontmatter that describes behaviour an agent should adopt. This project provides:

Crate	Description
`skill`	Core library — discovery, parsing, installation, lock-file management, provider registry
`skills-cli`	Binary (`skills`) — interactive CLI with the same UX as the TypeScript original

The Rust port achieves 100 % command parity with the original TypeScript CLI while adding the performance and reliability benefits of a native compiled binary.

Features

39 supported agents — Cursor, Claude Code, Windsurf, Cline, Codex, Roo Code, GitHub Copilot, Kilo Code, and many more
All source types — GitHub shorthand (owner/repo), GitHub / GitLab URLs with branch + subpath, local paths, well-known HTTP endpoints, direct git URLs
Plugin manifests — .claude-plugin/marketplace.json and plugin.json discovery for Claude Code ecosystem compatibility
Symlink-first install — canonical storage + per-agent symlinks (or junctions on Windows); copy mode available
Lock files — global (~/.agents/.skill-lock.json) and project-scoped (skills-lock.json) for reproducible setups
Extensible providers — HostProvider trait for adding custom skill hosts beyond GitHub / GitLab / well-known
Telemetry-aware — opt-in anonymous telemetry (respects DO_NOT_TRACK / DISABLE_TELEMETRY)
Cross-platform — Linux, macOS, Windows (with native junction support)

Source Formats

# GitHub shorthand (owner/repo)
skills add qntx/skills

# Full GitHub URL
skills add https://github.com/qntx/skills

# Direct path to a skill in a repo
skills add https://github.com/qntx/skills/tree/main/skills/code-review

# GitLab URL
skills add https://gitlab.com/org/repo

# Prefix shorthand
skills add github:owner/repo
skills add gitlab:owner/repo

# Local path
skills add ./my-local-skills
skills add /absolute/path/to/skills

# Well-known HTTP endpoints
skills add https://example.com   # checks /.well-known/skills/index.json

`skills add` Options

Option	Description
`-g, --global`	Install to user directory instead of project
`-a, --agent <agents...>`	Target specific agents (e.g., `claude-code`, `cursor`). Use `'*'` for all
`-s, --skill <skills...>`	Install specific skills by name (use `'*'` for all skills)
`-l, --list`	List available skills without installing
`--copy`	Copy files instead of symlinking to agent directories
`-y, --yes`	Skip all confirmation prompts
`--all`	Install all skills to all agents without prompts (shorthand for `--skill '' --agent '' -y`)

Examples

# List skills in a repository
skills add qntx/skills --list

# Install specific skills
skills add qntx/skills --skill frontend-design --skill code-review

# Install a skill with spaces in the name (must be quoted)
skills add owner/repo --skill "Convex Best Practices"

# Install to specific agents
skills add qntx/skills -a claude-code -a cursor

# Non-interactive installation (CI/CD friendly)
skills add qntx/skills --skill frontend-design -g -a claude-code -y

# Install all skills from a repo to all agents
skills add qntx/skills --all

# Install all skills to specific agents
skills add qntx/skills --skill '*' -a claude-code

# Install specific skills to all agents
skills add qntx/skills --agent '*' --skill frontend-design

Installation Scope

Scope	Flag	Location	Use Case
Project	(default)	`./<agent>/skills/`	Committed with your project, shared with team
Global	`-g`	`~/<agent>/skills/`	Available across all projects

Installation Methods

Method	Description
Symlink (Recommended)	Creates symlinks from each agent to a canonical copy. Single source of truth, easy updates.
Copy	Creates independent copies for each agent. Use when symlinks aren't supported.

Note: On Windows, junctions are used instead of symlinks for directory linking.

Other Commands

Command	Aliases	Description
`skills add <source>`	`a`, `install`, `i`	Add a skill package from any source
`skills remove <name>`	`rm`, `r`	Remove installed skills
`skills list`	`ls`	List installed skills
`skills find <query>`	`f`, `s`, `search`	Search the skills registry
`skills check`		Check for available skill updates
`skills update`	`upgrade`	Update all skills to latest versions
`skills init`		Scaffold a new `SKILL.md`
`skills experimental-install`		Restore skills from `skills-lock.json`
`skills experimental-sync`		Sync skills from `node_modules`

`skills list`

List all installed skills. Similar to npm ls.

# List all installed skills (project and global)
skills list

# List only global skills
skills ls -g

# Filter by specific agents
skills ls -a claude-code -a cursor

`skills find`

Search for skills interactively or by keyword.

# Interactive search (fzf-style)
skills find

# Search by keyword
skills find typescript

`skills check` / `skills update`

# Check if any installed skills have updates
skills check

# Update all skills to latest versions
skills update

`skills init`

# Create SKILL.md in current directory
skills init

# Create a new skill in a subdirectory
skills init my-skill

`skills remove`

Remove installed skills from agents.

# Remove interactively (select from installed skills)
skills remove

# Remove specific skill by name
skills remove web-design-guidelines

# Remove multiple skills
skills remove frontend-design web-design-guidelines

# Remove from global scope
skills remove --global web-design-guidelines

# Remove from specific agents only
skills remove --agent claude-code cursor my-skill

# Remove all installed skills without confirmation
skills remove --all

# Remove all skills from a specific agent
skills remove --skill '*' -a cursor

# Remove a specific skill from all agents
skills remove my-skill --agent '*'

# Use 'rm' alias
skills rm my-skill

Option	Description
`-g, --global`	Remove from global scope (~/) instead of project
`-a, --agent`	Remove from specific agents (use `'*'` for all)
`-s, --skill`	Specify skills to remove (use `'*'` for all)
`-y, --yes`	Skip confirmation prompts
`--all`	Shorthand for `--skill '' --agent '' -y`

Library Usage

use skill::SkillManager;
use skill::types::{DiscoverOptions, InstallOptions, InstallScope, InstallMode, AgentId};

#[tokio::main]
async fn main() -> skill::Result<()> {
    let manager = SkillManager::builder().build();

    // Discover skills in a directory
    let skills = manager
        .discover_skills(std::path::Path::new("./my-repo"), &DiscoverOptions::default())
        .await?;

    // Install a skill for a specific agent
    let opts = InstallOptions {
        scope: InstallScope::Project,
        mode: InstallMode::Symlink,
        cwd: None,
    };
    for skill in &skills {
        manager.install_skill(skill, &AgentId::new("cursor"), &opts).await?;
    }

    // List installed skills
    let installed = manager.list_installed(&Default::default()).await?;
    println!("Installed: {}", installed.len());

    Ok(())
}

Supported Agents

Click to expand the full list of 39 supported agents

Agent	Skills Dir	Global Skills Dir
Amp	`.agents/skills`	`$XDG_CONFIG_HOME/agents/skills`
Antigravity	`.agent/skills`	`~/.gemini/antigravity/skills`
Augment	`.augment/skills`	`~/.augment/skills`
Claude Code	`.claude/skills`	`~/.claude/skills`
OpenClaw	`.agents/skills`	`~/.openclaw/skills`
Cline	`.agents/skills`	`~/.agents/skills`
CodeBuddy	`.codebuddy/skills`	`~/.codebuddy/skills`
Codex	`.agents/skills`	`$CODEX_HOME/skills`
Command Code	`.commandcode/skills`	`~/.commandcode/skills`
Continue	`.continue/skills`	`~/.continue/skills`
Cortex Code	`.cortex/skills`	`~/.snowflake/cortex/skills`
Crush	`.crush/skills`	`~/.crush/skills`
Cursor	`.cursor/rules/skills`	`~/.cursor/rules/skills`
Droid	`.droid/skills`	`~/.droid/skills`
Factory Code	`.factory/skills`	`~/.factory/skills`
Gemini CLI	`.gemini/skills`	`~/.gemini/skills`
GitHub Copilot	`.github/skills`	`~/.github/skills`
Goose	`.goose/skills`	`~/.config/goose/skills`
iFlow CLI	`.iflow/skills`	`~/.iflow/skills`
Junie	`.junie/skills`	`~/.junie/skills`
Kilo Code	`.kilocode/skills`	`~/.kilocode/skills`
Kimi K2	`.kimi/skills`	`~/.kimi/skills`
Kiro	`.kiro/skills`	`~/.kiro/skills`
Kode	`.kode/skills`	`~/.kode/skills`
MCPJam	`.mcpjam/skills`	`~/.mcpjam/skills`
Mistral Vibe	`.vibe/skills`	`~/.vibe/skills`
Mux	`.mux/skills`	`~/.mux/skills`
Neovate	`.neovate/skills`	`~/.neovate/skills`
OpenCode	`.opencode/skills`	`~/.opencode/skills`
OpenHands	`.openhands/skills`	`~/.openhands/skills`
Pi	`.pi/skills`	`~/.pi/skills`
Qoder	`.qoder/skills`	`~/.qoder/skills`
Qwen Code	`.qwen/skills`	`~/.qwen/skills`
Replit	`.agents/skills`	`$XDG_CONFIG_HOME/agents/skills`
Roo Code	`.roo/skills`	`~/.roo/skills`
Trae	`.trae/skills`	`~/.trae/skills`
Trae CN	`.trae/skills`	`~/.trae-cn/skills`
Windsurf	`.windsurf/skills`	`~/.windsurf/skills`
ZenCoder	`.zencoder/skills`	`~/.zencoder/skills`

Creating Skills

Skills are directories containing a SKILL.md file with YAML frontmatter:

---
name: my-skill
description: What this skill does and when to use it
---

# My Skill

Instructions for the agent to follow when this skill is activated.

## When to Use

Describe the scenarios where this skill should be used.

## Steps

1. First, do this
2. Then, do that

Required Fields

name — Unique identifier (lowercase, hyphens allowed)
description — Brief explanation of what the skill does

Optional Fields

metadata.internal — Set to true to hide the skill from normal discovery. Internal skills are only visible when INSTALL_INTERNAL_SKILLS=1 is set.
metadata.tags — Array of tags for categorization and search.

---
name: my-internal-skill
description: An internal skill not shown by default
metadata:
  internal: true
  tags: [internal, wip]
---

Skill Discovery

The CLI searches for skills in these locations within a repository:

Root directory (if it contains SKILL.md)
skills/, skills/.curated/, skills/.experimental/, skills/.system/
Agent-specific directories: .claude/skills/, .cursor/skills/, .windsurf/skills/, etc.

Plugin Manifest Discovery

If .claude-plugin/marketplace.json or .claude-plugin/plugin.json exists, skills declared in those files are also discovered:

{
  "metadata": { "pluginRoot": "./plugins" },
  "plugins": [
    {
      "name": "my-plugin",
      "source": "my-plugin",
      "skills": ["./skills/review", "./skills/test"]
    }
  ]
}

This enables compatibility with the Claude Code plugin marketplace ecosystem.

Design Principles

Feature parity — Every command, source format, agent, and behaviour from the Vercel TypeScript implementation is faithfully reproduced.
Zero runtime dependencies — Single static binary; no Node.js, npm, or npx required.
Library-first — The skill crate exposes a clean SkillManager API so agent frameworks can embed skill support directly.
Security by default — Path traversal protection on subpaths and skill names; plugin manifest paths validated against base directory.
Extensible — Custom agents via AgentRegistry::register(), custom providers via the HostProvider trait, feature-gated network and telemetry.

Feature Flags

Flag	Default	Description
`network`	off	Enables `reqwest`-based network operations (GitHub API, well-known fetching)
`telemetry`	off	Enables anonymous usage telemetry to `https://add-skill.vercel.sh/t`

The CLI binary enables both flags. Library consumers can opt in selectively.

Environment Variables

Variable	Description
`CODEX_HOME`	Override Codex config directory (default: `~/.codex`)
`CLAUDE_CONFIG_DIR`	Override Claude Code config directory (default: `~/.claude`)
`XDG_CONFIG_HOME`	Override XDG config base (default: `~/.config`)
`INSTALL_INTERNAL_SKILLS`	Set to `1` or `true` to include internal skills
`GITHUB_TOKEN` / `GH_TOKEN`	GitHub API token for update checks (falls back to `gh auth token`)
`SKILLS_API_URL`	Override the skills search API endpoint (default: `https://skills.sh`)
`DISABLE_TELEMETRY` / `DO_NOT_TRACK`	Set to `1` or `true` to disable telemetry

Compatibility

This project is a drop-in replacement for the Vercel skills CLI (npx skills).

Same CLI interface, flags, and aliases
Same SKILL.md format and YAML frontmatter schema
Same lock-file format (skill-lock.json v3, skills-lock.json v1)
Same agent directory structures and detection logic
Same source parsing rules (GitHub, GitLab, local, well-known, git)
Same telemetry endpoint and event schema
Same installation semantics (canonical + symlink/junction)

Agent Compatibility

Skills are generally compatible across agents since they follow a shared Agent Skills specification. However, some features may be agent-specific:

Feature	Claude Code	Cursor	Windsurf	Codex	Cline	Roo Code	OpenCode
Basic skills	✅	✅	✅	✅	✅	✅	✅
`allowed-tools`	✅	✅	✅	✅	✅	✅	✅
`context: fork`	✅	❌	❌	❌	❌	❌	❌
Hooks	✅	❌	❌	❌	✅	❌	❌

Troubleshooting

"No skills found"

Ensure the repository contains valid SKILL.md files with both name and description in the frontmatter.

Skill not loading in agent

Verify the skill was installed to the correct path
Check the agent's documentation for skill loading requirements
Ensure the SKILL.md frontmatter is valid YAML

Permission errors

Ensure you have write access to the target directory. On Windows, you may need to run as Administrator for junction creation.

Symlink issues on Windows

Windows requires Developer Mode or Administrator privileges for symlinks. The CLI automatically falls back to junctions for directories.

Security

Path traversal protection — All subpaths and skill names are validated to prevent directory escape attacks.
Plugin manifest sandboxing — Paths declared in .claude-plugin/ manifests must be contained within the base directory and start with ./.
No arbitrary code execution — Skills are passive markdown instruction files; they do not execute code during installation.
Dependency audit — Minimal dependency tree; only well-known crates from the Rust ecosystem.

License

Licensed under either of:

Apache License, Version 2.0 (LICENSE-APACHE or https://www.apache.org/licenses/LICENSE-2.0)
MIT License (LICENSE-MIT or https://opensource.org/licenses/MIT)

at your option.

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in this project shall be dual-licensed as above, without any additional terms or conditions.

A QNTX open-source project.

Code is law. We write both.

Name		Name	Last commit message	Last commit date
Latest commit History 19 Commits
.github		.github
skill		skill
skills-cli		skills-cli
.gitattributes		.gitattributes
.gitignore		.gitignore
Cargo.lock		Cargo.lock
Cargo.toml		Cargo.toml
LICENSE-APACHE		LICENSE-APACHE
LICENSE-MIT		LICENSE-MIT
Makefile		Makefile
README.md		README.md
SECURITY.md		SECURITY.md
clippy.toml		clippy.toml
install.ps1		install.ps1
install.sh		install.sh
rust-toolchain.toml		rust-toolchain.toml
rustfmt.toml		rustfmt.toml

Folders and files

Latest commit

History

Repository files navigation

Skill

Install a Skill

Install the CLI

Overview

Features

Source Formats

skills add Options

Examples

Installation Scope

Installation Methods

Other Commands

skills list

skills find

skills check / skills update

skills init

skills remove

Library Usage

Supported Agents

Creating Skills

Required Fields

Optional Fields

Skill Discovery

Plugin Manifest Discovery

Design Principles

Feature Flags

Environment Variables

Compatibility

Agent Compatibility

Troubleshooting

"No skills found"

Skill not loading in agent

Permission errors

Symlink issues on Windows

Security

Related Links

License

About

Topics

Resources

License

Licenses found

Code of conduct

Contributing

Security policy

Uh oh!

Stars

Watchers

Forks

Releases 6

Contributors

Uh oh!

Languages

`skills add` Options

`skills list`

`skills find`

`skills check` / `skills update`

`skills init`

`skills remove`