Skip to main content

What Agent Skills do

Agent Skills are open, ready-to-use skills that help AI assistants produce high-quality, standards-aligned K-12 teaching materials. They are grounded in learning science and leverage Knowledge Graph. Each skill packages the instructions, references, and guardrails an agent needs to reliably complete a teacher workflow. With Agent Skills, the same task can produce consistent, classroom-ready results.
SkillDescription
K-12 Lesson PlanningBuilds classroom-ready, standards-aligned lesson plans, optionally aligned to a teacher’s curriculum.
K-12 Lesson DifferentiationAdapts an existing lesson into tiered versions (below, at, or above proficiency level) and for specific student needs, keeping core content consistent across tiers.

When to use Agent Skills

Use caseDescriptionExamples
Plan standards-aligned lessonsGenerate classroom-ready lesson plans with teacher-facing plans, student materials, and observation templates.A 3rd grade teacher needs a lesson on rounding to the nearest hundred for tomorrow.
Differentiate existing lessonsAdapt a lesson into below / at / above proficiency tiers while preserving core content and grade-level demand.A 6th grade teacher wants tiered versions of a food webs lesson linked from an external curriculum site.
Ground outputs in trusted dataPair Agent Skills with Knowledge Graph to resolve standards, learning components, progressions, and curriculum context.Anchor a math lesson to state standards and Illustrative Mathematics structures.
Benchmark instructional qualityScore generated materials against published eval rubrics using LLM-as-judge or human review.Validate that a lesson plan meets pedagogy, rigor, formatting, and model-scaffolding criteria.

How to access Agent Skills

Agent Skills are cross-platform and model-agnostic. They can be used with any agent runtime that supports the open Agent Skills format ↗:
  • Claude Code
  • Cursor
  • Codex
Check out the GitHub repository ↗ to inspect skill instructions, reference files, scripts, and rubrics directly.

Our approach

Our initial set of Agent Skills and evaluator rubrics was co-developed with Anthropic ↗. These skills don’t just generate plausible-looking materials — they’re grounded in research-backed principles and refined with expert practitioners.
PrincipleDescription
PedagogyOutputs reflect sound instructional design.
  • Standards alignment
  • Prerequisite and forward connections
  • Appropriate instructional model
  • Discourse structures
  • Attention to student struggle
  • Visual and representational choices
RigorOutputs maintain grade-level cognitive demand
UsabilityOutputs are classroom-ready and default to universal design