general-oat-skills

Runs a quick project health check covering git status, recent commits, GitHub Actions workflow runs, and package build verification. Use when starting a session, before a release, or to get a fast overview of project state.

Runs a complete code quality pipeline: ruff linting and formatting, mypy type checking, and pytest test suite. Use for end-of-session quality checks, pre-release validation, or when the user asks to lint, format, or clean up code.

Loads all markdown files from the current directory or a specified directory into context for analysis and reference. Non-recursive, loads only top-level .md files. Use when the user wants to read all docs in a folder, review documentation, or bring markdown content into the conversation.

Prepares and triggers a new Python package release: runs cleanup, validates semantic versioning, updates pyproject.toml and docs, builds with uv, creates a GitHub release, and monitors the publishing workflow. Supports patch/minor/major auto-increment. Use when the user wants to cut a release, bump the version, or publish a new Python package version.

Verify and release a Rust crate or workspace. Bumps version, commits, tags, pushes. CI-driven by default; --local publishes directly.

Run a comprehensive multi-agent QA audit of any project. Auto-detects language and framework, then launches parallel agents to audit code quality, architecture, error handling, test coverage, CI/CD, documentation, and security. Use for pre-release validation, periodic quality reviews, or when the user asks for a full project audit.

Runs Google Agent Development Kit (ADK) evaluations in sorted order with detailed results, failure summaries, and rate limiting between tests. Use when the user wants to run ADK evals, test agent performance, or validate agent behavior against evaluation datasets.

Creates a new project from templates with full setup including directory structure, dependency management, git initialization, CI/CD, testing, and documentation. Supports Python, Node.js, Rust, Go, and 14 AI agent framework templates. Use when the user wants to start a new project, bootstrap a codebase, or create a project from a template.

Interrogates a vague task one question at a time until it is fully scoped, then writes a ready-to-use prompt: what (goals), why (the problem), constraints (guardrails), done (success criteria), and the open questions it deliberately left unresolved, each with the assumption being made. Works for any task, not just code — writing, research, a decision, a process — building an analysis tree of the problem and asking only the questions whose answers are not already discoverable. Use when the user has a half-formed task, wants to turn an idea into a brief for Claude or Codex, or asks to scope, spec, or define a piece of work before starting it.

Interactive lesson-based teaching skill. Researches a topic via the web, builds a structured curriculum of up to 10 lessons, and delivers them one at a time with comprehension questions. Supports difficulty flags (--beginner, --advanced, --expert) and lesson count (--lessons N). Use when the user wants to learn about any topic in a guided, incremental way.

Runs the comprehensive test suite with pytest, reports coverage statistics, analyzes failures, and attempts automatic fixes. Use when the user wants to run tests, check coverage, or validate that code changes pass the test suite.