The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted content from external websites using tools like web_fetch and curl. This process creates a vulnerability surface where malicious instructions could be embedded in the HTML or metadata of the pages being audited.
Ingestion points: External web content fetched via web_fetch and curl (SKILL.md).
Boundary markers: The skill provides a clear warning to treat fetched pages as untrusted data and explicitly instructs the agent to ignore any instructions found in the page copy or meta tags (SKILL.md).
Capability inventory: The skill is primarily analytical and generates an audit report; it does not contain code for local file modification or arbitrary command execution.
Sanitization: The instructions note that web_fetch strips <script> tags, which helps mitigate the risk of script-based attacks during analysis.

seo-audit