The Agent Skills Directory

[SAFE]: The skill provides clear, specific instructions for SEO auditing without any signs of malicious intent or behavior. It correctly identifies canonical tags, robots directives, and site health issues through standard parsing.
[PROMPT_INJECTION]: The skill processes external web content (indirect prompt injection surface), but the risk is minimized by its deterministic logic and narrow scope of automated fixes.
Ingestion points: Processes rendered_dom, raw_html, and response headers from external URLs via WebFetch in SKILL.md.
Boundary markers: No specific delimiters or "ignore instructions" warnings are defined for the ingested content.
Capability inventory: Utilizes WebFetch for data retrieval and Bash for auditing and applying fixes as defined in the frontmatter and Audits section.
Sanitization: Automated fixes are limited to the injection or removal of specific HTML tags (<link rel="canonical"> and noindex) using deterministic logic, which limits the potential for arbitrary code execution from malicious web content.

seo-indexability