holoscan-setup

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

• Third-party content exposure detected (high risk: 0.75). SKILL.md:36 requires fetching public web content from https://docs.nvidia.com/... at runtime, which is outsider-authored free text that can be ingested into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs fetching https://docs.nvidia.com/holoscan/sdk-user-guide/ (and its sdk_installation.html page) at runtime to determine supported platforms, package names, and install requirements, so remote content will directly control the agent's installation recommendations.