garethmanning

education-agent-skills: Evidence-Grounded Skills for Education AI

A library of 165 structured, evidence-grounded skills for education tasks, designed for AI agents and manual use. It supports teachers, school leaders, curriculum designers, and EdTech builders across planning, assessment, learning science, and student study support.

TypeScriptOther licenseEducationAI AgentsAgent Skills
View on GitHub Compare
Stars
820
Forks
119
Stars, last 16 days
+62

Our take

Health score

43/100

How it is scored
Activity
34
Community
13
Issues
65
Pull requests
60

+62 stars in the last 16 days

  • Fewer contributors than 86% of the projects we track

Overview

education-agent-skills is a collection of structured prompts that help AI agents apply named educational research to teaching, curriculum, assessment, and learning support. It aims to offer a more research-grounded alternative to generic education prompts, while marking evidence strength and identifying approaches the project excludes.

The skills are designed for both individual use and orchestration: educators can copy a skill into an AI tool, while developers can use typed inputs, outputs, and chaining metadata to integrate skills into workflows. The first 19 domains focus on educators and designers; a separate domain provides live interaction patterns for students.

Key Features

  • Provides 165 skills across 20 education domains, from retrieval practice and curriculum design to inclusive design and historical thinking.
  • Includes YAML metadata with evidence ratings, named sources, typed input and output schemas, and chaining details.
  • Supports Claude, OpenAI Codex, Hermes Agent, Agent Skills-compatible tools, and manual copy-paste use.
  • Includes student-facing patterns intended to guide AI responses during study sessions, such as progressive hints and confidence checks.
  • Offers MCP tools for skill discovery and details, alongside prompts for the full skill collection.
  • Documents evidence limitations and excluded approaches, and labels original practitioner frameworks separately.

Use Cases

  • A teacher planning a unit can combine skills for backwards design, spaced practice, and formative assessment.
  • A curriculum team can use the library to review learning progressions, assessment alignment, or inclusive design barriers.
  • An EdTech developer can integrate structured education knowledge into an agent workflow using skill schemas and chaining metadata.
  • A student-facing AI study tool can use the domain 20 interaction patterns to structure hints, retrieval, and reflection.

Getting Started

For a local Codex setup, clone the repository and add it as a plugin marketplace:

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

For other tools and manual use, see the README.

Considerations

  • The prompts encode research-grounded practices, but the README notes that the prompts themselves have not been empirically validated as AI interventions.
  • Hosted MCP access requires an authorization token. Local installation and manual use remain available without it.
  • The repository README advertises CC BY-SA 4.0 for its educational materials, but the supplied repository metadata reports the license as NOASSERTION. Confirm the applicable terms before reuse.
  • The skills are a knowledge and prompt library, not a standalone teaching platform. They require an AI tool for interactive use, or can be copied and used manually.

Found this useful?

Share it with someone who would like education-agent-skills.