eljulians / skillfile
One-stop shop for AI skills and agents. Search 110K+ community skills, install and track them declaratively, and deploy across all major AI coding tools (Claude Code, Codex, etc)
AI Architecture Analysis
This repository is indexed by RepoMind. By analyzing eljulians/skillfile in our AI interface, you can instantly generate complete architecture diagrams, visualize control flows, and perform automated security audits across the entire codebase.
Our Agentic Context Augmented Generation (Agentic CAG) engine loads full source files into context on-demand, avoiding the fragmentation of traditional RAG systems. Ask questions about the architecture, dependencies, or specific features to see it in action.
Repository Overview (README excerpt)
Crawler viewskillfile [ ]() One-stop shop (and file) for AI skills and agents. Search, install, and track them declaratively across all major AI coding tools. Customize without losing upstream updates. Community skills and agents are popping up everywhere (agentskill.sh, skills.sh, GitHub repos, raw URLs). Installing them usually means one-liners, copy-pasting markdown, or running tool-specific plugins. Nothing tracks what you installed, there's no lock file, no way to update, and if you tweak a skill you lose your changes the next time you reinstall. gives you a single config file ( ) that declares everything. Run to browse 110K+ community skills right from your terminal! Including popularity and security scores. Or add entries by hand for repos you already know. fetches your skills and agents, locks them to exact commit SHAs, and deploys them where your AI tools expect them (7 platforms supported, including Claude Code or Codex). Edit an installed skill? captures your changes as a patch so they survive upstream updates. You stay in sync with the source without losing your customizations. Not a framework. Does not run agents. Just manages the markdown files that frameworks consume. Install From crates.io Pre-built binaries Download from GitHub Releases. From source Run for the full command list. Quick Start On a fresh clone, reads and fetches the exact pinned content -- fully reproducible. > **Note:** skillfile uses the GitHub API to resolve commits. Without a token you'll hit the 60 requests/hour rate limit fast. Set , , or run before using it. See Environment Variables for details. Key Features • **Declarative manifest** — one config file declares all your skills, agents, and platform targets • **Reproducible installs** — every entry pinned to an exact commit SHA via • **Cross-platform deploy** — one manifest, seven AI platforms • **Pinning & patching** — customize installed skills without losing upstream updates • **Search & discovery** — find skills across multiple registries from the CLI Skillfile Format Line-oriented, space-delimited, human-editable. No YAML, no TOML. Names are inferred from filename stems when omitted. See SPEC.md for the full format specification. | Field | Description | |---|---| | | Source type: , , | | | or | | | Logical name (inferred from filename if omitted). Must match . | | | (github) GitHub repository identifier | | | Path to the file (local: relative to repo root, github: within the repo) | | | (github) Branch, tag, or commit SHA. Defaults to . | | | (url) Direct URL to raw markdown file | Lock File resolves every GitHub entry to an exact commit SHA and writes it to . Commit this file — on a fresh clone, fetches the exact same bytes. re-resolves to the latest SHA upstream. Your lock file diffs cleanly in code review, so you always see what changed. Pinning & Patching Edit an installed skill, then it to survive upstream updates: Patches are stored in and committed to version control. Search & Discovery Find skills and agents across multiple registries without leaving the terminal: Searches agentskill.sh, skills.sh, and skillhub.club in parallel. Results are sorted by popularity and deduplicated. Supported Platforms | Platform | Skills directory | Agents directory | Scopes | |---|---|---|---| | | / | / | local, global | | | / | / | local, global | | | / | -- | local, global | | | / | / | local, global | | | / | -- | local, global | | | / | / | local, global | | | / | / | local, global | Multiple platforms can be configured simultaneously. Each line in the Skillfile adds a deployment target. Directory Layout Security Skillfile is a file manager. It downloads content from sources you specify and places it where your AI tools expect it. It does not analyze, verify, or sandbox the content it manages. The lock file pins entries to exact commit SHAs, giving you reproducibility -- the same SHA always produces the same bytes. lets you review what will be fetched. Patches make all local modifications visible in version control. But none of this tells you whether the content is safe to use. Review what you install. The risk profile is the same as . Environment Variables | Variable | Description | |---|---| | / | GitHub API token for SHA resolution and private repos. **Strongly recommended** -- unauthenticated requests hit GitHub's API rate limit (60 req/hour) very quickly. Set a token or run to get 5,000 req/hour. | | | Merge tool for | | | Fallback editor for | Contributing