The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze content from third-party websites provided as URL arguments, which presents a risk of processing embedded malicious instructions.
Ingestion points: The skill accepts a <url> argument for most primary commands, including /seo audit, /seo page, /seo technical, and /seo content (defined in SKILL.md).
Boundary markers: There are no specific instructions or boundary delimiters provided to the agent to treat fetched content as untrusted or to ignore instructions embedded within the HTML or metadata of the target URLs.
Capability inventory: The skill orchestrates a wide range of analysis capabilities, including technical auditing, E-E-A-T evaluation, and schema validation across multiple subagents (defined in SKILL.md).
Sanitization: The skill lacks explicit sanitization or filtering logic for the external content before it is processed by the AI for report generation.
[EXTERNAL_DOWNLOADS]: The reference documentation includes instructions for using external performance auditing tools.
The file references/cwv-thresholds.md recommends using the Lighthouse CLI via npx lighthouse, which fetches the package from the npm registry to perform local performance audits.

seo