The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external website content, representing a surface for indirect prompt injection. This is mitigated by explicit instructions for the agent to treat fetched data as untrusted and to ignore any commands found within it.\n
Ingestion points: The skill utilizes web_fetch and curl to ingest HTML and metadata from external sites.\n
Boundary markers: Includes specific warnings to disregard instructions found in page copy or meta tags.\n
Capability inventory: Uses network-capable tools and performs local file reads for marketing context.\n
Sanitization: Employs web_fetch, which filters out script tags from the processed output.\n- [DATA_EXPOSURE_EXFILTRATION]: The skill performs network operations to audit sites and accesses local configuration files.\n
Network operations: Uses curl and web_fetch to retrieve content from external domains as part of the audit process.\n
Local file access: Accesses .agents/product-marketing.md to inform SEO strategy.\n
External references: Links provided in reference files point to reputable SEO industry sources like Google Search Central and Yoast.\n- [COMMAND_EXECUTION]: The skill instructions reference the curl utility to fetch site data, which is appropriate for the skill's diagnostic functions.

seo-audit