ai-seo

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs the agent to use third‑party MCPs (DataForSEO, Semrush, SE Ranking) and to "research each tool (pricing, features, reviews)" and source pain points from public review sites like G2/Capterra, so the agent fetches and interprets untrusted, user‑generated web content that can change content‑generation and publishing actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly adds runtime MCP integrations that fetch and run remote code or data — e.g., cloning and running the GitHub repo https://github.com/AminForou/mcp-gsc (git clone ... && python ...), and connecting to the DataForSEO SSE endpoint https://mcp.dataforseo.com/sse — both are used at runtime and the fetched content can execute code or directly control agent behavior.