SKILL.md packages that extend Claude Code, Cursor, Copilot, and other AI agents.
Tags
decepticon
Techniques to probe and exploit LLM APIs: rate-limit abuse, token-cost amplification, schema bypass, model-version manipulation, and related probes.