seo
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data (website HTML) for SEO analysis, creating a risk of indirect prompt injection where hidden instructions in the website's content could affect agent behavior.
- Ingestion points: Website HTML and metadata retrieved during SEO audits (referenced in
skill-report.jsontemplates). - Boundary markers: Absent; there are no instructions to the agent to treat audited HTML content strictly as data or to ignore embedded instructions.
- Capability inventory: The skill is intended for use with agents capable of reading external content and providing technical analysis, which involves high-level reasoning over potentially malicious input.
- Sanitization: No sanitization or validation of the input HTML is defined.
Audit Metadata