The Agent Skills Directory

[SAFE]: The skill uses specific MCP tools (run_site_audit, get_backlinks_overview, etc.) for its intended purpose of SEO analysis. All external resources and tool naming conventions are consistent with the provided vendor context.
[SAFE]: The workflow includes a verification step where findings must be checked against live HTML before reporting, reducing the risk of hallucinated or malicious tool outputs being passed to the user.
[SAFE]: The report generation uses a local template.html file. The instructions explicitly warn against using dramatic language or jargon, favoring a plain-language, non-technical approach which minimizes the risk of deceptive metadata or social engineering.
[SAFE]: The skill includes an 'adversarial pass' step, instructing the agent to self-review or use a second model to check for jargon, overwhelming information, or claims beyond the verified facts, which serves as a safety guardrail.

seo-audit