perplexity-core-workflow-b

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime path takes outsider-provided topic / generated questions and sends them as messages to the Perplexity Sonar API (via perplexity.chat.completions.create using baseURL: "https://api.perplexity.ai"), so attacker-authored free text can be ingested as prompt content at selection time without the agent choosing a specific pre-authored item.