hunt-idor

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The content is a highly actionable offensive playbook describing deliberate exploitation techniques (header/tenant injection, token prediction, JWT/key minting, webhook/callback exfiltration, GraphQL pivots, SCIM path-body override, mass-assignment and agent-key minting) that enable credential theft, cross-tenant data exfiltration, and account takeover rather than merely benign diagnostics.

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs the agent/hunter to fetch and interpret arbitrary public third‑party content (e.g., crawl JS bundles and Swagger/OpenAPI specs like /swagger.json, introspect GraphQL endpoints, probe arbitrary URLs, and inspect mobile app HTTPS traffic via Frida/mitmproxy), and then use those responses to drive further tool actions, so untrusted content could supply indirect prompt-injection instructions that change behavior.