customer-research

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

• Third-party content exposure detected (high risk: 0.95). The skill’s runtime workflow uses UnifAPI to fetch live public community content (e.g., Reddit threads/comments via seo/serp → reddit/posts/{id}/comments, plus TikTok comments), which is outsider-authored free text that is ingested into the agent for synthesis.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). This skill explicitly depends on UnifAPI at runtime (https://unifapi.com/mcp) and calls its endpoints (e.g., seo/serp, reddit/posts/{id}/comments, tiktok/videos/{id}/comments) to fetch live public content that is injected into the agent context and thus directly controls the agent's prompts/outputs.