mistica

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

• Third-party content exposure detected (high risk: 0.70). The skill requires the agent to read documentation files and explicitly instructs a fallback to fetch the main llms.md from the public GitHub repo (https://github.com/Telefonica/mistica-web/blob/master/doc/llms.md), meaning it ingests open/public third‑party content that can directly influence its behavior.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs the agent to fetch and read the external documentation at https://github.com/Telefonica/mistica-web/blob/master/doc/llms.md at runtime when local node_modules aren’t present, and that fetched doc is treated as the canonical source that directly controls the agent’s prompts/instructions.