Packaged prompt engineering for Claude Code and friends

What it does

This repository bundles AI agent instructions into discrete, installable units following the Agent Skills format. Each unit contains a machine-facing SKILL.md with YAML frontmatter, a human-facing README.md, and optional helper scripts. The collection spans code analysis, documentation, architecture diagrams, dependency management, and workplace communication.

The interesting bit

Instead of treating prompts as throwaway chat history, the project packages them as reusable skills that slot into Claude Code, Codex, Cursor, and AdaL. It also provides specialized sub-agents and slash commands for delegated workflows—like codebase pattern matching or PR summarization—rather than single-shot completions.

Key highlights

• Over 40 skills across development, documentation, planning, testing, and professional workflows
• Installable as individual plugins across multiple AI coding agents (Claude Code, Codex, Cursor, AdaL)
• Includes Claude Code-exclusive sub-agents and slash commands for multi-step task delegation
• Each skill ships with both SKILL.md instructions for the agent and README.md docs for the user
• Optional scripts/ and references/ directories for automation helpers and supporting context

Caveats

• Agents and slash commands are exclusive to the Claude Code CLI and unavailable in other editors
• The collection is explicitly described as “opinionated,” reflecting one team’s workflows rather than universal standards
• Skills are individual plugins; there is no monolithic bundle beyond marketplace registration

Verdict

Useful if you spend your day in Claude Code and want structured, repeatable prompts for specialized tasks. Skip it if you need runtime libraries or framework code—this is pure instruction engineering.