# CLAUDE.md - PHP AI Coding Agent Context & Token Boundary Rules for `appwrite/appwrite`
[![PulseForge 2026 Empirical Benchmark](https://exactbaseline.com/api/badge/appwrite/appwrite.svg?tier=community)](https://exactbaseline.com/leaderboard?repo=appwrite/appwrite)

> **Generated by [PulseForge AI Empirical Radar](https://exactbaseline.com/leaderboard)** • Maintainer: `@appwrite` • Stack: `PHP` (62,000 KB / ~1,488,000 LOC)
> **Upgrade to Drop-In Migration & 45% Cost-Cut Config Pack ($4.99)**: https://exactbaseline.com/buy/micro_pass?repo=appwrite/appwrite
> **Complete 9-File AI Stack BOM & Repo Hardening .ZIP ($9.99)**: https://exactbaseline.com/buy/repo_pass_999?repo=appwrite/appwrite

## 1. Context Window & Token Boundary Rules (-58% Discovery Token Bloat)
- **Excluded Discovery Paths (NEVER read, index, or grep)**: `vendor/, composer.lock, storage/, bootstrap/cache/, .phpunit.result.cache`
- Before reading any file > 300 lines, query symbol definitions or line ranges first.
- Keep per-turn prompt context <= **26,880 tokens** (down from 64,000 unindexed tokens/turn; saves ~$433.11/mo).
- Keep every source file strictly under 1,000 lines.

## 2. PHP Architecture, Schema Boundary & Verification Gate for Appwrite
- **Verification Command**: Run `vendor/bin/phpunit && vendor/bin/phpstan analyse` after every contiguous code edit.
- **Schema & Boundary Discipline**: Strict types declare(strict_types=1) + PHPStan level-8 static analysis.
- Never hardcode API keys, bearer tokens, or `.env` secrets in source files or test fixtures.

## 3. Empirical Benchmark Attestation (63.5 -> 96.4/100)
- Live 4-Pillar Telemetry Profile: https://exactbaseline.com/pulseforge-leaderboard?repo=appwrite/appwrite
- Machine-Readable LLM Index: https://exactbaseline.com/pulseforge-leaderboard.md