Skip to content

AGENTS.md

Latest commit

 

History

History
34 lines (26 loc) · 1.73 KB

AGENTS.md

File metadata and controls

34 lines (26 loc) · 1.73 KB
Raw

This is NOT the Next.js you know

This version has breaking changes — APIs, conventions, and file structure may all differ from your training data. Read the relevant guide in node_modules/next/dist/docs/ before writing any code. Heed deprecation notices.

📘 Mandatory: Read Project Guidelines

BEFORE writing any code, you MUST read GUIDELINE.md in the project root.

This file contains:

  • Project architecture & folder structure conventions
  • Coding standards (TypeScript, React, Tailwind CSS v4)
  • Environment variable rules (.env, .env.development, .env.production)
  • PWA (Serwist) configuration guidelines
  • SEO & performance requirements
  • Git conventions
  • Dos and Don't for AI agents

Failure to follow these guidelines will result in inconsistent, non-maintainable code.

🧠 Mandatory: Skill Inspection & Invocation Before Execution

EVERY TIME the user gives any command or request, you MUST:

  1. Mandatory Skills to Always Invoke:
    • using-superpowers (.agents/skills/using-superpowers/SKILL.md): Read and follow BEFORE taking any action or responding (including clarifying questions).
    • ponytail (.agents/skills/ponytail/SKILL.md): Read and follow to enforce YAGNI, minimal code, zero unrequested abstractions, and maximum efficiency.
  2. Task-Specific Skills: Check all available skills in .agents/skills/ (and global skills). If any match the task, read their SKILL.md using view_file BEFORE writing code or taking action.
  3. Strictly follow all instructions defined in the matched skill(s).