DannyMac180/skills
AI agent skills created by me: Dan McAteer
What it solves
This repository provides a collection of reusable "skills" or plugins for AI agents, designed to extend their capabilities in workflow management, learning, and system modification. It addresses the need for structured, supervised agentic workflows and better control over how agents handle complex tasks.
How it works
Users can install these skills via the Claude Code plugin marketplace or by pointing an AI agent to the specific GitHub URL of a skill. The project includes several specialized tools:
- ModSmith: A tool for creating and managing Claude Mods, including token and prompt-cache review and vetting processes.
- Codex Dynamic Workflows: A system for planning and running supervised workflows using goal modes, subagents, and approval gates.
- Codex Relay: A mechanism for handing off tasks between native Codex tasks with compact handoffs.
- Explain This: A learning tool that creates a persistent learner profile to explain digital artifacts and provide comprehension quizzes.
Who it’s for
Developers and power users of AI agents (specifically Claude Code and Codex) who want to add advanced operational capabilities and ready-made behavioral modifications to their agents.
Highlights
- Supervised Workflows: Supports subagents, simulated work packets, and verification gates.
- Mod Management: Provides a way to vet and install Claude Mods with a focus on token and prompt-cache efficiency.
- Learning Integration: Includes a persistent learner profile for spaced-repetition review and customized explanations.
- Agent Agnostic: While optimized for Claude Code and Codex, it allows installation for any agent that supports skill-based extensions.
Related
- Project
- Project
- Project
- Project
- Project