fireflies

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs the agent to query Fireflies' GraphQL API (https://api.fireflies.ai/graphql) to fetch full meeting transcripts, sentences, summaries, and to upload audio from public HTTPS URLs—i.e., it ingests user-generated third-party meeting content that the agent will read and act on, so that content could carry embedded or indirect instructions.