Education Agent Skills: 165 AI Skills for Evidence-Based Teaching

Introduction

The education-agent-skills repository by GarethManning is an extensive open-source library featuring 165 evidence-based pedagogical AI skills. These skills are meticulously crafted for teachers, school leaders, and EdTech developers, covering critical areas such as pedagogy, learning science, curriculum design, assessment, and regenerative education. Engineered for AI agent orchestration, the library is compatible with platforms like Claude, OpenAI Codex, and Hermes Agent, providing a rigorous foundation for AI in educational contexts.

Each skill is grounded in named research, with transparent evidence strength ratings, ensuring that AI-powered educational tools are built on effective, proven practices rather than convention or assumption. The library is designed to support educators in creating personalized, evidence-grounded learning experiences and to streamline tasks like lesson planning and assessment design.

Installation

The library offers flexible installation and usage options, prioritizing sustainable free use through local installations.

Claude

• CoWork (Easiest): Go to Customize ? (+) Add Plugin and paste https://github.com/GarethManning/education-agent-skills.
• Claude Code CLI: Install from the repo URL: claude plugin install https://github.com/GarethManning/education-agent-skills.
• Claude.ai / Claude Desktop (Hosted MCP): For remote MCP endpoint needs, use https://mcp-server-sigma-sooty.vercel.app/mcp. Note that hosted access requires an access token.

OpenAI Codex

Codex does not require the hosted MCP server. The recommended local setup is:

git clone https://github.com/GarethManning/education-agent-skills.git
cd education-agent-skills
codex plugin marketplace add "$PWD"

Alternatively, copy individual skills into your global Codex skills directory: cp -r skills/<domain>/<skill-name> ~/.codex/skills/.

Hermes Agent

Hermes users can add the repository as a tap and install specific skills:

hermes skills tap add GarethManning/education-agent-skills
hermes skills install \
  GarethManning/education-agent-skills/skills/original-frameworks/learning-target-authoring-guide \
  --category education --yes

Any Agent Skills-compatible tool

Copy skill folders from the skills/ directory into your agent's skills directory. Each skill is a self-contained folder with SKILL.md.

Manual (No Setup)

1. Open any skill file in the repository (under skills/).
2. Copy the prompt block.
3. Paste it into any AI and fill in the fields for your class or context.

Examples

Here are practical examples of how to leverage the education-agent-skills library:

With a Runtime Install (Recommended)

Install the skills in Claude, Codex, or Hermes, then describe your needs in plain language. The agent can automatically select and run relevant skills.

Example: Tell your agent, "I'm planning a Year 9 science unit on cells, 6 weeks, 3 lessons a week."

Claude can then run the Backwards Design Unit Planner, Spaced Practice Scheduler, and Retrieval Practice Generator in parallel. In minutes, you receive a complete lesson-by-lesson plan, incorporating spaced retrieval, evidence-grounded sequencing, and ready-to-use formative assessment activities, all tailored to your specified timeline and topics.

Without the Plugin (Manual)

No API key, no technical setup, no dependencies.

Example: Open skills/memory-learning-science/spaced-practice-scheduler/SKILL.md and provide:

• Topics: Cell structure, Cell transport, Cell division, Enzymes, Biological molecules
• Timeline: 8-week term, starting 3 February
• Lessons per week: 3

Claude will return a complete week-by-week schedule. This schedule outlines when to teach new content and when to revisit previous topics at expanding intervals, including specific retrieval activities for each review slot. The schedule adheres to research on optimal spacing intervals and interleaving, providing practical guidance for addressing learning gaps.

Why Use It

This library stands out due to its commitment to evidence-based practice and its practical design for educators and developers alike:

• Evidence-Driven: Every skill is rooted in named research, transparently rated for evidence strength (Strong, Moderate, Emerging, Original). Frameworks lacking empirical support, such as learning styles, are explicitly excluded. This ensures that AI applications in education are built on what truly works.
• Built by an Educator: Developed by Gareth Manning, an educator with 20 years of international school experience, the pedagogical judgments embedded in each prompt reflect real classroom and curriculum design practice.
• Designed for Orchestration: With YAML schema headers, typed input/output fields, and chaining metadata, the skills are engineered for programmatic use and seamless integration into complex AI agent workflows.
• Comprehensive Domains: The library covers 20 diverse domains, from Memory & Learning Science to AI Literacy and Student-Facing Learning Skills, offering a holistic approach to educational AI.
• Empowers Educators: It provides structured, evidence-grounded support in minutes, freeing up teachers to focus on the human aspects of teaching. It also offers a credible foundation for personalized learning, curriculum audits, and professional development.

Links

• GitHub Repository: https://github.com/GarethManning/education-agent-skills
• License: CC BY-SA 4.0
• Contact Gareth Manning:
  • Email: gareth.manning@gmail.com
  • X (formerly Twitter): https://x.com/worldteacherman
  • LinkedIn: https://www.linkedin.com/in/gareth-manning-a404b387/