Back to directory

Wednesday Agent Skills

A set of pre-configured skills and a pre-computed dependency graph that give AI coding agents structural knowledge of a specific codebase, plus git hooks that enforce complexity limits and design-system rules.

Wednesday Agent Skills

This system builds an AST-based SQLite dependency graph of a codebase once, mapping every import, export and function call across languages such as JavaScript, TypeScript, Python, Go and Swift. AI agents then answer structural questions, such as what a change would affect, from that pre-computed graph instead of re-reading source files.

It installs into Claude Code, Cursor, Gemini CLI, GitHub Copilot and other tools as an on-demand skill, and generates or updates each tool's own instruction file, such as CLAUDE.md or .cursorrules, with the available skills and project rules. It also installs git hooks that enforce a cyclomatic-complexity limit, require a purpose header on new files, enforce conventional commit messages, and restrict UI code to an approved set of components.

It installs with npx, is written in JavaScript, and is MIT licensed.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.