The Agent Skills Directory

[SAFE]: The skill performs legitimate SEO auditing functions such as inspecting HTML metadata, validating structured data, and checking crawlability. These operations are consistent with the skill's stated purpose and involve standard browser automation tools.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface due to its ingestion of external web content.
Ingestion points: Reading DOM content via inspect_page, fetching robots.txt, and parsing sitemap.xml.
Boundary markers: Absent.
Capability inventory: The agent writes reports and regression tests to the local shiplight/ directory.
Sanitization: Content is extracted for technical analysis against SEO standards rather than being interpreted as commands.

seo-review