hunt-misc

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (critical risk: 1.00). This skill contains explicit, actionable exploit techniques and payloads (SSRF token exfiltration, SAML signature stripping for account takeover, invitation/token misuse, CRLF/header injection, token-scope bypass) that directly instruct on taking over accounts and exfiltrating credentials.

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill contains multiple curl/code examples that embed tokens, session cookies, and API secrets directly in headers or request bodies (Authorization: token ghp_..., Cookie: session=..., PRIVATE-TOKEN, auth_token), which would force an LLM to insert real secret values verbatim when instantiating those examples.