llm-cost-optimization-auditlisted

# LLM Cost-Optimization Audit — $49 You added prompt caching and your bill went up. Caching a prefix you reuse less often than the TTL window is a net loss, not a no-op. Profiles first, then audits the levers that actually move an inference bill, each saving quoted with its precondition. **This is a preview.** The full skill (method, Iron Rules, dated snapshots, worked examples) is available on Agensi: [agensi.io/skills/llm-cost-optimization-audit](https://www.agensi.io/skills/llm-cost-optimization-audit)