thesys-c1-genui

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

• Third-party content exposure detected (high risk: 0.80). The skill includes an explicit webSearch tool (getWebSearchTool) that calls exa.search to fetch web search results and returns those results into the LLM workflow—i.e., it ingests open/public web content (third‑party, user-generated) for the agent to read and act on.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill makes runtime calls to the Thesys C1 API (e.g., https://api.thesys.dev/v1/embed and https://api.thesys.dev/v1/artifact) which return C1 DSL that directly controls the agent's UI/actions and is a required dependency for the skill, so this external content can influence agent behavior at runtime.