grad-cct

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

• Third-party content exposure detected (high risk: 1.00). The skill's methodology and example (SKILL.md Step 2 and examples/sample_scenario.md) explicitly instruct using netnography and "public community data"—naming Dcard, Facebook groups, and Instagram hashtags—as primary data sources, meaning the agent is expected to fetch and interpret untrusted, user-generated third-party social media content that can directly influence its analysis and recommendations.