Vibeleaderboard
← Back to Tools
Index / tool

spm (Skill Package Manager)

camunda.github.io
Visit camunda.github.io
Category
Developer Tools
Rank
No. 2276Tools index

Previous survey · No. 2175 ·

Pricing
Open Source
Type
TOOL
Use case
Agent Building
Interfaces
CLI
Date

About

A CLI tool that lets developers declare AI skills as git dependencies in an ai.json manifest, pinning them to tags, branches, or commits via a lockfile, then materializes them into vendor-specific locations for tools like Claude Code, Cursor, Copilot CLI, and Gemini CLI without ever committing the skill files to the repo. It uses a shared global cache and a single cross-platform binary that shells out to git for reproducible, node_modules-style installs.

What it does

A command-line utility that lets teams list AI assistant skills in a manifest file, lock each one to a specific git commit, and generate copies laid out for the file structures that different AI coding tools expect, using a shared local cache to avoid repeated downloads.

Stated on the product site

Version pinning
Skill versions are pinned to a tag, branch, or commit and locked to an immutable commit SHA in a lockfile.
Repo footprint
Materialized skill files are gitignored rather than committed to version control.
Supported tools
Projects skills into project-local locations for multiple AI coding tool vendors.
Distribution
Distributed as a single cross-platform binary that calls out to the system's git installation.
Caching
Uses a shared local cache so each repository at a given commit is fetched only once across projects.

Not stated on the site

  • The page does not specify a license under which the software is distributed.
  • The page does not describe any cost or pricing for using the tool.

Written from the product site at camunda.github.io.

What it can do

  • Declare AI skills as git dependencies in a manifest file

    ai.json manifest → Skill dependency declarations

  • Pin skill dependencies to specific tags, branches, or commits

    Git repository reference → ai.lock lockfile entry

  • Materialize declared skills into vendor-specific configuration locations

    ai.json/ai.lock skill definitions → Skill files placed in Claude Code, Cursor, Copilot CLI, or Gemini CLI config locations

  • Install skills by shelling out to git

    Git repository URL → Downloaded skill files

  • Cache installed skills in a shared global cache

    Downloaded skill package → Cached skill files reusable across projects

  • Keep skill files out of the project's version control

    Skill package → Project repository without committed skill files

Tags

clipackage-managerai-skillsdeveloper-toolsgitclaude-codegithub-copilot

Comments (0)

No comments yet

Editorially curated, with community endorsements as a secondary signal. Corrections welcome.