ai-seo
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill recommends several third-party SEO monitoring tools and services, including Otterly AI, Peec AI, ZipTie, and LLMrefs. It also suggests using
npx is-agentic, a website audit tool provided by Vercel (a recognized trusted organization). - [COMMAND_EXECUTION]: The instructions advise the user or agent to execute the shell command
npx is-agentic yourdomain.comto evaluate a site's readiness for AI agents. This is a standard procedure for the referenced tool. - [INDIRECT_PROMPT_INJECTION]: The skill guides the agent to perform audits by ingesting content from untrusted external sources (user domain, Wikipedia, Reddit, and YouTube) without defining boundary markers or sanitization requirements. The agent possesses capabilities such as shell command execution (
npx). This combination establishes a vulnerability surface for indirect prompt injection attacks.
Audit Metadata