learning-science 开源项目推荐（74 个）

nagisanzenin /engram

面向 Claude Code 的循证学习引擎——第一性原理课程、带凭据的自由回忆验证、FSRS 调度记忆和可探索工件。学习任何内容；保留所学。

GarethManning /education-agent-skills

165 evidence-grounded AI skills for teachers, school leaders and EdTech builders—pedagogy, learning science, curriculum, assessment and regeneration. Claude, Codex and Hermes.

24kchengYe /human-skill-tree

🌳 AI驱动的技能树，用于终身人类学习。从K-12到职业与社会智力的30+技能，基于认知科学。|人类养成记：AI 驱动的终身学习技能树

learning-commons-org /knowledge-graph

The data layer for smarter educational AI. Integrate trusted instructional content and research directly into your AI-powered tools — improving precision, relevance, and instructional alignment.

learning-commons-org /evaluators

Evaluation for AI outputs against trusted educational rubrics. Measure and improve content quality with research-backed rubrics — ensuring rigor, reliability, and alignment to classroom needs.

lowwwbank /anything-to-course

Universal AI agent skill that turns any material into learning-science-based self-study courses with retrieval practice, spaced repetition, and tutor mode.

Aisland-SJL /Aisland-teach

🎓 A teaching skill for AI coding agents — teaches from your level, then verifies real learning by restatement. Adapted from Matt Pocock's teach (MIT).

PSkinnerTech /brainlift-writer-skill

AI-agnostic skill for writing BrainLifts: structured context docs that turn research into DOK knowledge trees, insights, and spiky POVs.

remi-huang /cognitive-learning-canvas

把硬核长内容拆成obsidian的思维导图 · 通用 AI 提示词框架(Claude/ChatGPT/Gemini)

jarmolkowicz /knowledge-base-modrn-mind

Research-backed concepts for educators teaching AI literacy. How to preserve human thinking, judgment, and agency while using AI productively.

pipeshub-ai /sharp

Stay sharp while AI does the work: a Claude Code plugin for judgment, understanding and transfer — one-way door gates, change briefs, and lessons that come back when you need them.

fanfanfanfanfan626 /mastery-tutor

Local-first mastery tutor for AI agents with guided HTML lessons, hands-on practice, transfer checks, spaced review, and learner-owned progress.

yuanmingze2008 /human-aware-teaching-skill

An Agent Skill for AI teaching that reasons about the learner's current understanding and chooses the next teaching move, grounded in cognitive and learning science.

epic-coder97 /no-sidequests

Distraction-free AI learning. Stay with one concept at a time using clear explanations, useful analogies, visual structure, and active recall.

code-yeongyu /memory-systems-lab

Glossy interactive React lab for human and agent memory systems.

ruhmahashmi /AILA

An Intelligent Lecturing Assistant

Norma-Vault /recall-quiz

Agent Skill that quizzes you on its own answers — one multiple-choice question after each substantive reply, graded and explained. Ambient retrieval practice against AI deskilling. Instruction-only: no scripts, no network, no data sent. MIT.

davidme6 /ai-learning-method

Evidence-driven Agent Skill for systematic learning: retrieval, discrimination, transfer, performance, and long-term continuity.

HuanMoovo /fuxi

伏羲框架 · 万物皆可学 — Everything Can Be Learned｜十阶时间线 × 证据分级 × 开源工具链

Da-boss-is-here /Lexforge

Eleven skill dimensions per word, FSRS-6 scheduling per skill, and no mastery until you can produce the word in writing. Each mistake is sorted into one of seven error categories, and no word enters practice before it's taught. Runs in the browser from one HTML file, with no server, no account and no network requests