EFECT: Executive Function Enhancement for Cognitive Technology

TL;DR

• 66 rule cards you can inject into LLMs
• 17 research modules with evidence-based techniques
• Works locally, no cloud needed
• Quick start →

---

What Is This?

Open-source experimental tools for integrating evidence-based ADHD support techniques into LLM interactions. Supports local LLMs via open standards for data sovereignty.

Heads up: This is a work-in-progress prototype. Content was compiled from literature sources with LLM assistance and needs manual verification. Currently built for tinkering, learning, and exploring. Obviously not a substitute for professional care.

---

Quick Start

I want to...

Browse techniques → Check out research/README.md
Build something → See rule_cards/ for structured cards
Test things → Read study_logs/README.md for tracking

Paths

🔍 Just Curious?

1. Check out research/README.md to see what's in here
2. Browse the modules 01-10g—they cover everything from CBT to mindfulness to cognitive training
3. Each module has techniques, effect sizes, and sources you can verify (please do!)

🛠️ Want to Build Something?

1. Look at rule_cards/ to see how techniques are structured
2. Check rule_cards/SPEC_RuleCard_v1.mdc for the format
3. See scripts/trigger.ps1 for an example implementation

🧪 Testing Things Out?

1. Read study_logs/README.md for how to log what you're doing
2. Use the card selection (see scripts/README.md)

---

What's Inside

• Evidence-Based: Real techniques largely automatically compiled from actual meta-analyses and reviews
• Modular: 17+ modules covering CBT, mindfulness, cognitive training, and more
• Rule Cards: Structured .mdc files you can drop into your LLM setup
• Reproducible: Selection so you can actually test things, and see what works for you

---

Rule Cards

The rule_cards/ folder has structured .mdc files you can inject into your LLM context.

The idea: Pick a card, inject it into the LLM's system prompt or context window, and the LLM follows that technique's protocol.

What Each Card Contains

• What it is: Metadata, sources, effect rating
• When to use it: Triggers
• How to do it: Protocol steps
• Constraints and rules: Non-negotiable limits
• What to track: Metrics

How to Use

Basic setup:

1. Pick a card (e.g., rc_stop_skill.mdc for the STOP technique)
2. Parse the .mdc file and format it as instructions
3. Inject it into your LLM's system prompt or prepend it to the conversation context
4. The LLM will follow the card's protocol when triggered

Example integration:

# Load card
card = load_mdc("rule_cards/rc_stop_skill.mdc")
# Format as system instruction
system_prompt = format_card_as_instructions(card)
# Inject into LLM context
llm.add_system_message(system_prompt)

Why Proper Integration Matters

Just pasting a card into chat doesn't work well because:

• Trigger detection: The LLM needs to recognize when to apply the card (trigger conditions). Rule based trigger is not stable / deterministic.
• Logging: Metrics and logging schemas need programmatic capture, not manual tracking

Proper integration means building a tool/UI that:

• Selects cards based on triggers or user choice / settings
• Injects cards into the right context layer (system vs. user)
• Monitors and enforces the protocol steps
• Tracks metrics and logs outcomes
• Manages context window according to card specifications

Check rule_cards/README.md for the full spec and scripts/trigger for an example implementation.

---

Documentation

The research/ folder has modular Markdown files covering non-pharmacological interventions for adult ADHD:

• 01-03: CBT basics (planning, problem-solving, psychoeducation)
• 04-06: Emotion stuff (DBT skills, mindfulness, cognitive training)
• 07-09: Coaching, workplace stuff, neurostimulation
• 10a-10g: Third-wave CBT, neurofeedback, exercise, sleep, nutrition, digital tools, HR checklists

Each Module Contains

• How to actually do the techniques
• Effect sizes (small/medium/etc.)
• Why it works (mechanisms)
• Sources you can (and should!) check (with page numbers!)
• Practical stuff for testing

More details in research/README.md.

---

Project Overview

EFECT helps you build tools that integrate evidence-based ADHD support techniques into LLM interactions. Think of it as a toolkit for making LLMs actually helpful for people with ADHD, instead of just another attention-sucking platform.

Two Main Parts

1. Documentation: Compiled evidence-based techniques from dozens of meta-analyses on ADHD—ready to use or remix
2. Implementation Examples: Working code and rule cards you can run locally, no cloud required

The idea: Help people with ADHD and attention challenges use LLMs productively, without getting lost in the attention economy's dark patterns.

---

Repository Structure

efect/
├── research/              # Evidence-based intervention modules (01-10g)
├── rule_cards/           # Structured rule cards for LLM integration
├── study_logs/           # Logging system for effectiveness tracking
└── scripts/             # Utility scripts for the system

---

Use It, But Use It Wisely

This is an experimental prototype - perfect for:

• Experimenting with new assistive tools
• Experimenting with LLM applications
• Learning how this stuff works
• Adjusting and improving

Things to Keep in Mind

• Verify everything: All data and citations need manual checking - don't trust, verify!!
• It's experimental: Use with care, test things
• Privacy matters: This stuff can be sensitive (like an LLM) and log sensitive data! (like an LLM) - handle data responsibly
• Not medical advice: This is a toolkit, not a replacement for professional care

You're responsible for what you build with this. Check sources, test thoroughly, and consult professionals when it matters.

---

Contributing

This is open source - change on it, improve it, remake it. Contributions, corrections, and wild ideas are all welcome.

See CONTRIBUTING.md for guidelines.

---

How This Was Made

This was compiled with heavy use of ChatGPT 5 thinking for literature discovery, drafting, and synthesis. All claims are theoretically backed by traceable sources.

Quality Status: This is a prototype. Sources have been auto-checked and spot-verified, but full manual review is needed. Verify citations yourself before using them for anything important. More details in research/README.md.

---

Note: This documentation is designed to enable others to develop similar systems and build upon this, while the implementations demonstrate practical applications.

📊 Coverage Report

✅ Rule Cards 0 / 66 verified Coverage: 0.0% Based on research_verified field

📚 Research Modules 0 / 139 verified Coverage: 0.0% Based on research_all_techniques.csv

---

📋 Module Breakdown

Module	Verified	Total	Progress
ADHS Coaching Skills	0	5	░░░░░░░░░░
Bewegung Sport Erwachsene ADHS	0	5	░░░░░░░░░░
CBT Emotionsregulation DBTSkills	0	20	░░░░░░░░░░
CBT Planen Organisieren	0	15	░░░░░░░░░░
CBT Problemlösen	0	15	░░░░░░░░░░
CBT Psychoedukation	0	15	░░░░░░░░░░
CBT ThirdWave integriert	0	6	░░░░░░░░░░
Digitale Formate iCBT Tele	0	5	░░░░░░░░░░
Ernährung Supplements ADHS	0	3	░░░░░░░░░░
HR Checkliste Arbeitsplatz ADHS erweitert	0	8	░░░░░░░░░░
Kognitive Remedation Training	0	9	░░░░░░░░░░
MBIs Achtsamkeit ohne CBT	0	11	░░░░░░░░░░
Neurofeedback Erwachsene ADHS	0	4	░░░░░░░░░░
Neurostimulation tDCS rTMS	0	5	░░░░░░░░░░
Psychoedukation Angehoerige Arbeitsumfeld	0	8	░░░░░░░░░░
Schlaf Interventionen ADHS Erwachsene	0	5	░░░░░░░░░░

---

License

This project is licensed under Apache License 2.0 - see LICENSE file for details.

---

Citation

If you use this documentation, please cite:

Adult ADHD Non-Pharmacological Intervention Modules (v0.1).
Compiled with extensive use of ChatGPT 5 (GPT-5 Thinking) for 
literature discovery and synthesis, with manual verification and 
source anchoring.