cargo-quickstart

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

• Third-party content exposure detected (high risk: 0.75). The required workflow uses curl to download workflow outputs from a signed URL (cargo-ai orchestration run download-outputs ... --format json), and those outputs include outsider-authored lead/job text from third-party connectors (e.g., Sales Navigator), which is then ingested into the LLM context via the agent’s subsequent “show the table”/reasoning over /tmp/quickstart-outputs.json.