boost-skill-cli

What it is and what it does

boost-skill-cli solves the problem of managing AI coding skills across multiple AI agents. A skill is a Markdown file with YAML frontmatter that gives an AI agent a repeatable capability—a workflow, house rules, or domain knowledge. Without a package manager, developers manually copy skill files into separate agent folders, losing track of versions and duplicating effort across teams. boost automates this: it taps GitHub-hosted registries, searches their contents with BM25 keyword matching (or optional semantic search), and installs chosen skills into all four major agents from a single lock file.

The default install is zero-dependency—pure Python stdlib with no build step. Search indexes build automatically on first use. Optional semantic search requires installing the [rag] extra, which downloads a small embedding model (~133 MB) on first search; you can also provide an API key for a larger model. The package maintains a curated catalog of 460+ registries across six domains (AI, architecture, UI, Java, ecommerce, infra), with measured item counts rather than estimates.

Use it for:

• Install a TDD workflow skill once and have it available in Claude Code, Cursor, Windsurf, and Gemini CLI simultaneously with version pinning.
• Search a curated registry of 460+ GitHub skill packs by keyword or meaning to find domain-specific rules or workflows for your tech stack.
• Export a catalogue as a 10.9 MB gzipped bundle and share it with teammates, who can import and search 59,972 items without cloning any repositories.
• Configure semantic search by installing the [rag] extra and running reindex to rank skill results by meaning instead of just keywords.
• Manage skills across multiple agents without manual file copying, keeping all instances version-synced in a single lock file.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

boost-skill-cli is a package manager for AI coding skills that finds, installs, and version-tracks skill files from GitHub registries, wiring them into Claude Code, Windsurf, Cursor, and Gemini CLI in a single operation.

Yes, if you actively use Claude Code, Windsurf, Cursor, or Gemini CLI and want to centrally manage and version-track AI coding skills across them. The zero-dependency default install and active maintenance make it low-risk. GPLv3 copyleft is a constraint only if you plan to redistribute modified versions. Not necessary if you use a single agent or manually manage skills.

Install

boost-skill-cli on PyPI

pip install boost-skill-cli

uv add boost-skill-cli

poetry add boost-skill-cli

Installing boost-skill-cli

Before you install

Low install friction: zero runtime dependencies, pure stdlib by default. Active maintenance with release 1 day old. Requires Python 3.12+.

License in practice

GPLv3 copyleft license means any modifications or derivative works must also be released under GPLv3. You can use and distribute the package freely, but cannot incorporate it into proprietary software without releasing your modifications under the same license.

Quickstart

pipx install boost-skill-cli
boost tap --defaults
boost search tdd
boost install tdd-workflow

Requires Python 3.12+ and git installed on the system.

Verify before relying

• Whether the BM25 search engine or semantic search (via [rag] extra) performs adequately for your skill discovery workflow.
• Compatibility and integration quality with each of Claude Code, Windsurf, Cursor, and Gemini CLI across different OS versions.
• Whether the 7 default registries and curated catalog meet your domain-specific skill needs.

Package facts

Evidence: boost_skill_cli-1.0.451-py3-none-any.whl

Keywords: ai, skills, package-manager, claude-code, cursor, windsurf, cli

Tags

More Build Tools packages

Further reading