hunt-captcha-bypass

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The document gives explicit, actionable instructions for bypassing CAPTCHA protections (omitting tokens, replaying tokens, concurrent requests to defeat rate-limits), which is guidance to weaken security controls and enable automated abuse.

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill explicitly instructs capturing a solved CAPTCHA token (e.g., g-recaptcha-response) and replaying the identical token in subsequent requests, which requires the agent to read and emit secret/token values verbatim in generated requests.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). SKILL.md describes a runtime workflow for CAPTCHA-bypass testing by instructing the agent to directly omit/replay CAPTCHA fields in HTTP POST requests to target endpoints, which involves ingesting outsider-supplied CAPTCHA/session form text from those requests.