← Documentation Home

For AI Assistants

This section is specifically designed for AI coding assistants (Claude, Kiro, Cursor, GitHub Copilot, etc.) working on the LLM Platform ecosystem.


🚨 CRITICAL - READ FIRST

Critical Warnings - Documented disasters from AI mistakes. Contains warnings about IRREVERSIBLE ACTIONS that have caused significant damage.

npm versions are PERMANENT - Once published, gone forever.
Issue closing requires verification - Read comments before closing.
Bulk operations need approval - Never assume state.

These warnings exist because of real disasters on 2025-11-30.


🚀 Start Here

If you're an AI assistant just starting to work on this codebase:

  1. Read First: AI Instructions - Critical rules and constraints
  2. Quick Reference: Command Reference - Essential commands at a glance
  3. Workflows: Workflow Rules - Mandatory development workflows
  4. Full Policies: Agent Policies - Complete policy documentation

📋 Quick Context

You're working on the LLM Ecosystem - an enterprise AI agent platform:

Component Technology
Primary CMS Drupal 11
Agent Standard OSSA v0.2.4
CLI Tooling BuildKit
Deployment Kubernetes
Source Control GitLab
CI/CD GitLab CI with merge trains

🚨 Critical Rules Summary

DO NOT:

ALWAYS:


📁 Key Paths

Purpose Path
Working Directory (read-only) ~/Sites/LLM/
Feature Work (worktrees) ~/Sites/LLM/worktrees/
Drupal Custom Code ~/Sites/LLM/all_drupal_custom/
NPM Packages ~/Sites/LLM/common_npm/
This Documentation ~/Sites/LLM/technical-documentation/
API Schemas ~/Sites/LLM/api-schema-registry/
Tokens ~/Sites/LLM/.env.local or ~/.tokens/

🔧 Essential Commands

# Before starting work
buildkit golden optimize         # Check if agents needed

# Testing
buildkit golden test             # Full test suite

# Drupal sync
buildkit drupal sync             # Sync custom code

# Quality
buildkit drupal phpcs PATH --fix # Fix PHP code style

# Git worktree
git worktree add ~/Sites/LLM/worktrees/<branch> <branch>

📚 Document Index

Document Purpose
🚨 Critical Warnings READ FIRST - Irreversible action safeguards
AI Instructions Primary onboarding, critical rules
Quick Reference Command cheatsheet
Workflow Rules Issue-first, worktree, TDD, npm publishing workflows
Agent Policies Complete policy documentation


Remember: When in doubt, ask the user. Don't assume, don't guess, don't create issues without approval.