seo-page-sections
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONOBFUSCATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [OBFUSCATION]: The skill instructions (Rule 5) explicitly direct the agent to remove characters such as em-dashes and en-dashes from its output to prevent identification by AI content detection filters.
- [COMMAND_EXECUTION]: The Python script
scripts/section_audit.py, which the agent is directed to run, configures its fetching logic to ignore SSL certificate validation (usingssl.CERT_NONE), making all network connections to external websites vulnerable to man-in-the-middle attacks. - [INDIRECT_PROMPT_INJECTION]: The skill ingests raw HTML content from arbitrary third-party URLs provided by the user. It processes this data without clear boundary markers or sanitization, potentially exposing the agent to instructions hidden within the target website's content.
Audit Metadata