Skip to main content

Skill Manager - manage agent skills from GitHub repos

Project description

SKM - Skill Manager

A CLI tool that manages global AI agent skills from GitHub repos or local directories. Clone repos or link local paths, detect skills via SKILL.md, and symlink them into agent directories — all driven by a single YAML config.

Note: skm manages skills at the user level (e.g. ~/.claude/skills/), not at the project level. It is not intended for installing skills into project-scoped directories.

skm install

Install

uv tool install git+https://github.com/reorx/skm

Quick Start

  1. Create ~/.config/skm/skills.yaml:
packages:
  - repo: https://github.com/vercel-labs/agent-skills
    skills:
      - vercel-react-best-practices
      - vercel-react-native-skills
  - repo: https://github.com/blader/humanizer
  - local_path: ~/Code/my-custom-skills       # use a local directory instead of a git repo
  1. Run install (skm i for short):
skm install

Skills are cloned (or symlinked from local paths) into your agent directories (~/.claude/skills/, ~/.codex/skills/, etc.).

Install from a source directly

You can also install skills directly from a repo URL or local path — no need to edit skills.yaml first:

# Install from a GitHub repo (interactive skill & agent selection)
skm install https://github.com/vercel-labs/agent-skills

# Install a specific skill by name
skm install https://github.com/vercel-labs/agent-skills vercel-react-best-practices

# Install from a local directory
skm install ~/Code/my-custom-skills

# Skip interactive prompts with --agents-includes / --agents-excludes
skm install https://github.com/blader/humanizer --agents-includes claude,codex

This detects available skills, lets you pick which ones to install (unless a specific skill name is given), and automatically adds the package to your skills.yaml config.

Commands

Command Description
skm install (or skm i) Install all packages from config. Clone repos (or link local paths), detect skills, symlink to agents, write lock file. Idempotent — also removes stale links (see below).
skm install <source> [skill] (or skm i) Install directly from a repo URL or local path without editing config. Interactively select skills and agents, then auto-update config.
skm list Show installed skills and their linked paths.
skm list --all Show all skills across all agent directories, marking which are managed by skm.
skm view <source> Browse and preview skills from a repo URL or local path without installing.
skm check-updates Fetch remotes and show available updates (skips local_path packages).
skm update <skill> Pull latest for a skill's repo, re-detect, re-link, update lock (skips local_path packages).

Config Format

~/.config/skm/skills.yaml:

agents:
  default:                   # optional: select which agents are active (omit = all)
    - claude
    - standard

packages:
  - repo: https://github.com/vercel-labs/agent-skills
    skills:                  # optional: install only these skills (omit = all)
      - vercel-react-best-practices
    agents:                  # optional: further filter agents for this package
      excludes:
        - standard

  - repo: https://github.com/blader/humanizer   # installs all detected skills to default agents

  - local_path: ~/Code/my-custom-skills         # use a local directory as package source
    skills:
      - my-skill

Each package must specify exactly one of repo or local_path. Local path packages use the directory directly (no cloning) and are skipped by check-updates and update.

Install Sync Behavior

skm install treats skills.yaml as a declarative state file. Each run syncs the agent directories to match the config:

  • New skills are linked to agent directories.
  • Removed skills (dropped from a package's skills: list, or entire package removed) have their links deleted.
  • Agent config changes (e.g. adding excludes: [openclaw]) remove links from excluded agents while keeping links in others.

Only links tracked in skills-lock.yaml are affected. Manually created files or skills installed by other tools in agent directories are never touched.

Skill Detection

A skill is a directory containing a SKILL.md file with YAML frontmatter (name field required). Detection order:

  1. Root SKILL.md — the repo itself is a singleton skill
  2. ./skills/ subdirectory exists — scan its children
  3. Otherwise — walk all subdirectories from repo root
  4. Stop descending once SKILL.md is found (no nested skills)

Known Agents

Skills are symlinked into these directories by default:

Agent Path
standard ~/.agents/skills/
claude ~/.claude/skills/
codex ~/.codex/skills/
openclaw ~/.openclaw/skills/

CLI Path Overrides

Override default paths for testing or custom setups:

skm --config /tmp/test.yaml \
    --store /tmp/store \
    --lock /tmp/lock.yaml \
    --agents-dir /tmp/agents \
    install

Key Paths

  • Config: ~/.config/skm/skills.yaml
  • Lock: ~/.config/skm/skills-lock.yaml
  • Store: ~/.local/share/skm/skills/

Development

uv sync
uv run pytest -v      # run tests
uv run skm --help     # run CLI

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

skm_cli-0.1.0.tar.gz (867.7 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

skm_cli-0.1.0-py3-none-any.whl (24.8 kB view details)

Uploaded Python 3

File details

Details for the file skm_cli-0.1.0.tar.gz.

File metadata

  • Download URL: skm_cli-0.1.0.tar.gz
  • Upload date:
  • Size: 867.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.9.15 {"installer":{"name":"uv","version":"0.9.15","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"macOS","version":null,"id":null,"libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":null}

File hashes

Hashes for skm_cli-0.1.0.tar.gz
Algorithm Hash digest
SHA256 c7ae3ef53415e3c79125507eac5b7b5bf03970e5857488504e38635dd5bea30d
MD5 9b081f66956884bbffbdb70cbc45c461
BLAKE2b-256 058c0bfd22820e35ec9effbd914d8d171e4ce82a0ccdd8f1cbe0f5e018a04907

See more details on using hashes here.

File details

Details for the file skm_cli-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: skm_cli-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 24.8 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.9.15 {"installer":{"name":"uv","version":"0.9.15","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"macOS","version":null,"id":null,"libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":null}

File hashes

Hashes for skm_cli-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 4c35d64b61cf7f5052d3b76171e810a64e0a93e8972667bffb0e06e301cdf668
MD5 65ce14ddd634d364cadda803c45efb00
BLAKE2b-256 08cd3d8d9c3d81fafc0738e2a395c73349de1203ce7b40eab29f935a5095470a

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page