seo-audit
Pass
Audited by Gen Agent Trust Hub on May 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates within its intended scope of search engine optimization. It guides the agent to perform structured audits without attempting to bypass safety protocols or access unauthorized data.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a process for fetching and analyzing external website content via
web_fetchandcurl. This creates a theoretical surface for indirect prompt injection where instructions embedded in a target site could influence the agent, but this is a necessary feature for SEO analysis and is considered a managed risk in this context. - Ingestion points:
SKILL.md(via operations fetching content from external URLs). - Boundary markers: None defined; the agent is instructed to read the raw output of web fetches.
- Capability inventory: Reading local configuration files (
product-marketing-context.md), executing shell-based web requests (curl), and utilizing web fetching tools. - Sanitization: No explicit sanitization or filtering of fetched HTML content is mandated beyond structural analysis.
Audit Metadata