The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because its core functionality depends on ingesting untrusted content from external websites during an audit. \n- Ingestion points: Data from target URLs is retrieved using tools like 'web_fetch' and 'curl' as referenced in the Technical SEO Audit section of SKILL.md. \n- Boundary markers: The instructions do not specify the use of delimiters or provide explicit warnings for the agent to disregard instructions found within the fetched HTML. \n- Capability inventory: The skill allows the agent to make network requests and read local project context files (e.g., product-marketing-context.md). \n- Sanitization: SKILL.md notes that the 'web_fetch' tool automatically strips script tags, which acts as a basic filter for external data. \n- [SAFE]: All external URLs and references in the documentation lead to reputable and well-known SEO industry sources, official developer platforms, and trusted organizations.

seo-audit