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

> **Generated by [PulseForge AI Empirical Radar](https://exactbaseline.com/leaderboard)** • Maintainer: `@papercupsio` • Stack: `Elixir` (4,816 KB / ~115,584 LOC)
> **Upgrade to Drop-In Migration & 45% Cost-Cut Config Pack ($4.99)**: https://exactbaseline.com/buy/micro_pass?repo=papercups-io/papercups
> **Complete 9-File AI Stack BOM & Repo Hardening .ZIP ($9.99)**: https://exactbaseline.com/buy/repo_pass_999?repo=papercups-io/papercups

## 1. Context Window & Token Boundary Rules (-58% Discovery Token Bloat)
- **Excluded Discovery Paths (NEVER read, index, or grep)**: `_build/, deps/, mix.lock, cover/, priv/static/assets/`
- Before reading any file > 300 lines, query symbol definitions or line ranges first.
- Keep per-turn prompt context <= **4,854 tokens** (down from 11,558 unindexed tokens/turn; saves ~$78.22/mo).
- Keep every source file strictly under 1,000 lines.

## 2. Elixir Architecture, Schema Boundary & Verification Gate for Papercups
- **Verification Command**: Run `mix test && mix credo --strict` after every contiguous code edit.
- **Schema & Boundary Discipline**: Ecto.Changeset strict cast/validate + Dialyzer typespec verification.
- Never hardcode API keys, bearer tokens, or `.env` secrets in source files or test fixtures.

## 3. Empirical Benchmark Attestation (60.2 -> 97.2/100)
- Live 4-Pillar Telemetry Profile: https://exactbaseline.com/pulseforge-leaderboard?repo=papercups-io/papercups
- Machine-Readable LLM Index: https://exactbaseline.com/pulseforge-leaderboard.md