Team-gohypemedia

optimize codex, claude code, cursor, copilot, and other ai coding-agent workflows for lower token usage without sacrificing correctness. use when users ask to reduce token consumption, implement caveman-style replies, compress coding prompts, clean or create agents.md, set file-selection guardrails, prevent full-file rewrites, route work across models, or diagnose why coding agents burn context or hit usage limits.