deep-research

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

• Third-party content exposure detected (high risk: 0.90). The skill's Step 3 "Evidence Collection (Deepresearch Tool)" instructs the agent to collect and preserve raw tool output and build an evidence table with "URL or reference", i.e., ingesting external/public web sources (including lower-tier opinion/aggregator sources), so it clearly reads and interprets untrusted third-party content as part of its workflow.