me2resh

Agent Decision Records (AgDR) - A standard for documenting technical decisions made by AI coding agents. Extends ADR for the age of AI-assisted development.

23
2
100% credibility
Found Feb 09, 2026 at 20 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Shell
AI Summary

This repository offers a standard format and integration guides for AI coding assistants to automatically document their technical decisions in human-readable records stored with the project code.

How It Works

1
📰 Discover AgDR

While using an AI helper to build your project, you learn about a simple way to make the AI explain its big choices so everyone understands.

2
📁 Add Guides to Your Project

You copy friendly notes into your project's folder to remind your AI helper to document decisions.

3
Pick Your AI Helper
🧠
Claude or Cursor

Your AI gets a special note to pause and record choices before building.

💻
Copilot or Others

Add reminders in the right spots so the AI always thinks twice and writes it down.

4
💭 Ask AI to Decide

You tell your AI, 'Choose the best way to handle logins,' and it thoughtfully compares options and creates a clear explanation.

5
📄 Review the Record

A new note appears in your project's docs folder, explaining the choice, pros, cons, and why it was picked.

6
👥 Share with Team

Your team reads the note to understand the AI's reasoning, making reviews easy and knowledge stick around.

🎉 Clear, Traceable Decisions

Now every AI choice is documented forever, helping your project stay organized and everyone on the same page.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 20 to 23 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is agent-decision-record?

Agent Decision Records (AgDR) standardizes documenting technical choices made by AI coding agents like GitHub Copilot, Claude Code, Cursor, and Codex, extending the classic ADR format for ai agent decision making. It captures why an agent picked Vitest over Jest or JWT for auth—context, options, tradeoffs—in markdown files stored per-project in docs/agdr/. Written in Shell, it provides git hooks, system prompts, and tool-specific plugins to enforce creation via commands like /decide or $agdr-decide.

Why is it gaining traction?

In the age of agent github copilot and decisions copilot agent, AI decisions vanish without trace, breaking audits and onboarding; AgDR makes them visible and searchable like human ADRs. Integrations with agent github action workflows, pre-commit hooks, and prompts for Cursor or Windsurf automate enforcement, turning vague "use React" into structured records with pros/cons tables. Developers notice faster team reviews and better ai agent decision tree traceability without manual overhead.

Who should use this?

Backend teams relying on GitHub Copilot or Claude Code for library picks and architecture in Node.js or TypeScript monorepos. AI-first consultancies auditing agent github code outputs. Frontend devs using Cursor for pattern decisions, tired of unexplained PR changes from multi agent decision transformer setups.

Verdict

Worth adopting early in AI-heavy projects for its solid docs and plug-and-play integrations, despite 21 stars and 1.0% credibility score signaling low maturity—no tests or broad adoption yet. Fork and contribute if agent decisionnaire fits your workflow; otherwise, stick to plain ADRs.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.