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

> **Generated by [PulseForge AI Empirical Radar](https://exactbaseline.com/leaderboard)** • Maintainer: `@qdrant_engine` • Stack: `Rust` (18,400 KB / ~441,600 LOC)
> **Upgrade to Drop-In Migration & 45% Cost-Cut Config Pack ($4.99)**: https://exactbaseline.com/buy/micro_pass?repo=qdrant/qdrant
> **Complete 9-File AI Stack BOM & Repo Hardening .ZIP ($9.99)**: https://exactbaseline.com/buy/repo_pass_999?repo=qdrant/qdrant

## 1. Context Window & Token Boundary Rules (-58% Discovery Token Bloat)
- **Excluded Discovery Paths (NEVER read, index, or grep)**: `target/, Cargo.lock, benches/fixtures/, *.rlib, *.rmeta`
- Before reading any file > 300 lines, query symbol definitions or line ranges first.
- Keep per-turn prompt context <= **18,547 tokens** (down from 44,160 unindexed tokens/turn; saves ~$298.85/mo).
- Keep every source file strictly under 1,000 lines.

## 2. Rust Architecture, Schema Boundary & Verification Gate for Qdrant
- **Verification Command**: Run `cargo test --workspace --locked && cargo clippy --all-targets -- -D warnings` after every contiguous code edit.
- **Schema & Boundary Discipline**: serde + schemars compile-time JSON Schema & zero-unwrap error boundaries.
- Never hardcode API keys, bearer tokens, or `.env` secrets in source files or test fixtures.

## 3. Empirical Benchmark Attestation (64.9 -> 96.0/100)
- Live 4-Pillar Telemetry Profile: https://exactbaseline.com/pulseforge-leaderboard?repo=qdrant/qdrant
- Machine-Readable LLM Index: https://exactbaseline.com/pulseforge-leaderboard.md