web-pentest

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt includes command examples that require inserting sensitive tokens verbatim (e.g., "$JWT", -H "Cookie: s=...") into generated commands/scripts, which would force an LLM to handle or output secrets directly and create an exfiltration risk.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The repository is an offensive security toolkit containing explicit, high-risk offensive features — credential exfiltration (SSRF to cloud metadata, OOB collaborator callbacks, JS cookie exfil hooks), authentication compromise (JWT forgery/bruteforce, jku/jwk/kid injection), and multiple remote-code-execution/persistence primitives (SSTI payloads, prototype-pollution NODE_OPTIONS --import data: URI gadget, gopher:// Redis packets that can write cron/SSH keys) plus request-smuggling and race-exploit tooling — all direct capabilities that can be used as backdoors or for unauthorized system compromise.