sidekick-ai

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

• Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md instructs the agent to use Membrane to run actions and proxy requests against the Sidekick AI API (e.g., "membrane action run" and "membrane request CONNECTION_ID /path/to/endpoint"), which will fetch user-generated meeting notes, messages, documents and other third-party content that the agent is expected to read and act on, exposing it to potential indirect prompt injection.