Show HN: A Claude Code skill that tells Claude when not to use Claude

Category: devtools

Tags: llm-audit, cost-optimization, agent-skill

Score: 6.8/10 (Innovation: 7, Technical: 6, Documentation: 8, Utility: 6)

This tool is a Claude Code skill that audits a codebase for LLM API calls, classifies them by task type, and recommends deterministic replacements where applicable. It's interesting because it addresses the growing problem of unnecessary LLM usage in projects by providing a structured audit and replacement workflow, potentially saving costs and complexity.

Target audience: software engineers, machine learning engineers, team leads

Repository: https://github.com/metravod/llm-buster-skill · MIT

View on Hacker News