competition-request-normalization-smuggling

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill explicitly requires capturing and preserving "raw request pairs", headers, and proxy/backend logs which can contain sensitive credentials (Authorization headers, cookies, tokens) and thus would force the LLM to handle or output secret values verbatim.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This repository explicitly instructs how to craft and replay HTTP request-smuggling differentials to reach unintended/hidden endpoints and bypass authentication, which is a high-risk offensive capability that can be used to compromise systems.