Imported from ActiveInferenceInstitute/courses (course_development/active_inference_ms/AGENTS.md). Install upstream with npx skills add ActiveInferenceInstitute/courses --skill active_inference_ms. Copyright stays with the author.

Active Inference for Middle School — Agent Guidelines

Quick Navigation: README | Resources | Real Life Skills | Body Science | Math Detectives | Code & Create | Portfolio AGENTS

Overview

This directory contains a 4-unit Active Inference curriculum for middle school students (grades 6-8) with 32 modules, shared resources, and comprehensive documentation. Agents working in this repository must maintain consistency in terminology, notation, and pedagogical structure while keeping content relatable, engaging, and connected to real-life experiences.

Directory Contents

Path	Type	Description
`README.md`	File	Curriculum overview, learning pathway, and navigation
`AGENTS.md`	File	This file — agent guidelines and content standards
`resources/`	Directory	Shared notation, glossary, references, and cross-course map
`01_real_life_skills/`	Directory	Unit 1: Real Life Skills — Active Inference in Daily Life (8 modules)
`02_body_science/`	Directory	Unit 2: Body Science — How Your Body Predicts (8 modules)
`03_math_detectives/`	Directory	Unit 3: Math Detectives — Patterns and Probability (8 modules)
`04_code_create/`	Directory	Unit 4: Code & Create — Building Smart Systems (8 modules)

Critical Rules for All Agents

1. Consult Shared Resources First

Before generating or editing any content, read these files:

Resource	Purpose	When to Consult
resources/notation_table.md	Canonical notation	Before writing any math or symbols
resources/glossary.md	Canonical definitions	Before using or defining any technical term
resources/references.md	Canonical references	Before citing any source
resources/cross_course_map.md	Cross-course links	Before adding cross-references

2. Never Use Placeholders

All content must use real methods — no mocks, stubs, placeholder brackets, or [TODO] markers. Every module must contain substantive, accurate content.

3. Maintain Unit-Specific Perspectives

Unit	Perspective	Lab Type	Example Content
Real Life Skills	Everyday intuition	Group Challenge	"Why does your phone's autocorrect sometimes guess wrong? What prediction did it make?"
Body Science	Biological mechanisms	Investigation Lab	"Test your reaction time — how does your brain predict when to catch a ball?"
Math Detectives	Pattern recognition & probability	Guided Worksheet	"Flip a coin 20 times. How close is your result to 50/50?"
Code & Create	Block coding & simple programming	Guided Coding Lab	"Use Scratch to build a creature that learns to avoid obstacles"

4. Write for Middle School Students (Grades 6-8)

Use relatable, real-world examples (social media, phones, sports, school)
Be curious and slightly irreverent — match the energy of a 12-year-old
Light math only: fractions, percentages, basic probability, ratios
Introduce coding through block-based tools (Scratch) before text-based
Define technical terms with everyday parallels
Use "you" and direct questions to keep engagement high

Notation Standards

All units use the notation defined in resources/notation_table.md. Key concepts:

Symbol/Concept	Kid-Friendly Version	Formal Meaning
Probability (%)	"How likely is it?"	P(event) expressed as percentage
50/50	"Equal chance either way"	P = 0.5
Prediction	"Your brain's best guess"	Expected value from generative model
Prediction error	"When your guess is wrong"	Difference between predicted and actual observation
Feedback loop	"When the result changes your next guess"	Recursive model updating
Pattern	"Something that repeats in a predictable way"	Statistical regularity

Terminology Standards

Preferred Term	Simpler Alternative Used First	Formal Meaning
Generative Model	Brain's prediction machine	Internal model that generates predictions
Prediction Error	Being wrong (and that's OK!)	Mismatch between expected and actual input
Markov Blanket	System boundary, bubble	Statistical boundary separating inside from outside
Precision	How sure you are	Inverse variance — weight on a signal
Free Energy	Overall wrongness score	Variational free energy (upper bound on surprisal)
Policy	Game plan, strategy	Sequence of actions
Epistemic	"Finding out"	Related to gaining information
Pragmatic	"Getting what you want"	Related to achieving goals

Topic Order Convention

All 4 units follow this exact topic order:

Systems → 2. Agents → 3. Perception → 4. Cognition → 5. Action → 6. Learning → 7. Communication → 8. Planning

Middle school version of the dependency chain:

What's a system? (Systems) → What makes you an agent? (Agents) → How do you notice stuff? (Perception) → How does your brain figure things out? (Cognition) → What do you decide to do? (Action) → How do you get better at stuff? (Learning) → How do you share what you know? (Communication) → How do you make plans? (Planning)

Content Format Standards

File	Format Requirements
`module.md`	`# Title: Subtitle` → Hook/Introduction → Learning Objectives → Key Vocabulary → Core Concepts (5 subsections with real-life examples) → Try This → Summary → References
`questions.md`	`# Course — Module — Study Questions` + simple numbered list (15-20 questions). Mix of factual and "what do you think?"
`practice_quiz.md`	`Name/Date` header → `Part A: Multiple Choice` (7 questions) → `Part B: Short Answer` (3 questions)
`lab.md`	`Objectives` → multi-part with `> Learning Goal:` blockquotes → `<!-- lab:reflection -->` / `{fill:textarea}` fields → `Summary` table
`dashboard.html`	Interactive HTML5: dark theme (`#2dd4bf` teal accent), concept cards with progress meters, quiz with JS answer checking
`README.md`	Quick Navigation header, overview, module contents table, cross-references, navigation footer
`AGENTS.md`	Quick Navigation header, directory contents table, content generation conventions

Cross-Reference Convention

When linking between units, use relative paths:

<!-- From 01_real_life_skills/03_perception/module.md to the body science version: -->
See [How Your Body Predicts](../../02_body_science/03_perception/module.md) for the science behind how your brain senses things.

<!-- From any module to the notation table: -->
See the [Notation Table](../../resources/notation_table.md) for symbol definitions.

<!-- From any module to the root README: -->
See the [Curriculum Overview](../../README.md) for the full course map.

Quality Checklist for Content Review

Before marking any module complete, verify:

Dashboard Color Identity

Accent: Teal #2dd4bf
Gradient: Teal → Blue
Semantic: Curious, explorative

active_inference_ms

About