mcollina/init
Creates, updates, or optimizes an AGENTS.md file for a repository with minimal, high-signal instructions covering non-discoverable coding conventions, tooling quirks, workflow preferences, and project-specific rules that agents cannot infer from reading the codebase. Use when setting up agent instructions or Claude configuration for a new repository, when an existing AGENTS.md is too long, generic, or stale, when agents repeatedly make avoidable mistakes, or when repository workflows have changed and the agent configuration needs pruning. Applies a discoverability filter—omitting anything Claude can learn from README, code, config, or directory structure—and a quality gate to verify each line remains accurate and operationally significant.
85
94%
Does it follow best practices?
Impact
72%
1.14xAverage score across 5 eval scenarios
Passed
No known issues
task.mdevals/scenario-3/
Our AGENTS.md has gotten out of hand. It's grown over time as different team members added things, and at this point it's just too long — a lot of it feels like stuff that's either obvious or that you could figure out just by looking at the repo. New engineers and agents alike tend to skip it entirely because it's so bloated.
Can you go through it and cut it down to just the parts that are actually worth keeping? We want something tighter that people will actually read and that gives real value — not a wall of generic advice.
The file is below. Please produce an improved version.
# Project Guide for AI Agents
## About this project
This is a Node.js web application built with Express.js and TypeScript. We use PostgreSQL for our database and Redis for caching. The frontend is React with TailwindCSS.
## Directory structure
- `src/` - Source code
- `src/api/` - API route handlers
- `src/models/` - Database models
- `src/utils/` - Utility functions
- `tests/` - Test files
- `scripts/` - Utility scripts
- `dist/` - Compiled output (don't edit)
## Code style
- Use 2-space indentation
- Use single quotes for strings
- Add semicolons at end of statements
- Use camelCase for variables and functions
- Use PascalCase for classes and interfaces
## Development workflow
- Write descriptive commit messages
- Test your code before committing
- Use feature branches and open PRs for review
- Keep PRs small and focused
## Commands
- `npm install` - Install dependencies (use --legacy-peer-deps flag as some peer deps have version conflicts)
- `npm run dev` - Start development server
- `npm run build` - Build for production
- `npm test` - Run tests
- `npm run lint` - Run ESLint
## Important notes
- Always run `scripts/migrate.sh` after pulling from main — it applies any pending DB migrations
- The `payments/` module has its own deployment pipeline separate from the main app — do not include payment changes in standard releases
- TypeScript strict mode is enabled
- ESLint and Prettier are configured — run lint before committing
- The database connection string is in `.env` — copy `.env.example` to get started
## Architecture
The app follows a layered architecture with controllers calling services which call repositories. The API layer handles HTTP concerns while business logic lives in services.