spydra

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly uses Spydra (a web-scraping tool) and instructs the agent to issue proxy requests to the Spydra API via Membrane (see "Spydra is a web scraping and data extraction tool" and "membrane request CONNECTION_ID /path/to/endpoint" in SKILL.md), which means the agent can ingest arbitrary public/third‑party web content that could contain embedded instructions.