What is SkillCatalog, a Git-native skill manager for AI coding tools?
SkillCatalog manages the instruction files that AI coding tools like Claude Code, Cursor, and Codex read from disk to shape their behavior. Hand-editing these files per-repo works for solo developers; it breaks down on a team of ten across multiple repositories. The app treats skills more like editor preferences than packages — there's a shared team baseline, but individual engineers layer their own selections on top without touching the shared catalog. The hierarchy is skills (Markdown with frontmatter) → stacks (ordered skill lists) → bundles (ordered stack lists) → profiles. Home profiles apply everywhere; project profiles inherit from them, with project-specific selections taking priority. Access control is Git access — no server to run, no external service, nothing leaves the machine unless you push. A drift detector flags hand-edited delivered files, which is useful when debugging unexpected AI behavior. The stack is a Rust Tauri v2 desktop app with a React frontend, plus a skc CLI for scripting and CI. Available on macOS via Homebrew; Linux and Windows are in progress.
What is SkillCatalog, a Git-native skill manager for AI coding tools used for?
Engineering teams use SkillCatalog to manage and version the AI coding-tool instruction files (Claude Code, Cursor, Codex) shared across multiple repositories using Git as the delivery layer.
How much does SkillCatalog, a Git-native skill manager for AI coding tools cost?
Pros and cons
Pros
• Treats skills as layered preferences — a shared team baseline plus per-engineer selections that don't touch the shared catalog • Clear hierarchy of skills → stacks → bundles → profiles, with project profiles inheriting from home profiles • Uses Git access as the access-control layer, so there's no server to run and nothing leaves the machine unless you push • Drift detector flags hand-edited delivered files, which helps when debugging unexpected AI behavior • Ships as a Rust Tauri desktop app plus an skc CLI for scripting and CI
Cons
• Desktop app is only available on macOS via Homebrew; Linux and Windows support is still in progress • Requires a team already using Git and file-based AI coding tools, so it's less useful for solo developers • No hosted service means each machine manages its own state rather than a central dashboard
Questions about SkillCatalog, a Git-native skill manager for AI coding tools
Which AI coding tools does SkillCatalog work with?
SkillCatalog manages the instruction files that file-based AI coding tools read from disk, including Claude Code, Cursor, and Codex. Skills are plain Markdown with frontmatter, so any tool that reads local instruction files can consume them.
Does SkillCatalog send my skill files to a server?
No. There's no server to run and no external service. Access control is handled through Git access, and nothing leaves your machine unless you choose to push it to a repository.
What is the drift detector for?
The drift detector flags delivered instruction files that have been hand-edited outside SkillCatalog. It's mainly useful when you're debugging unexpected AI behavior and need to see whether a file no longer matches the source skill.
Is SkillCatalog available on Windows or Linux?
Not yet. The desktop app is currently available on macOS through Homebrew, with Linux and Windows builds listed as in progress. The skc CLI is available for scripting and CI use.
How do team and individual skill selections coexist?
SkillCatalog keeps a shared team baseline while letting individual engineers layer their own selections on top without editing the shared catalog. Home profiles apply everywhere, and project profiles inherit from them with project-specific selections taking priority.
Reviews
Sign in to rate and review SkillCatalog, a Git-native skill manager for AI coding tools.
Sign in to reviewNo reviews yet. Be the first to review SkillCatalog, a Git-native skill manager for AI coding tools!