google-forums-search

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

• Third-party content exposure detected (high risk: 0.85). At runtime the script calls a public Frevana endpoint via curl, then reads the returned HTTP body from RESPONSE_FILE and prints it to stdout after only JSON validation (cat "$RESULT_FILE"), so any free-text contained in the API response is ingested into the agent’s LLM context.