The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill has a potential surface for indirect prompt injection as it is designed to ingest and process data from external websites.
Ingestion points: The skill instructions explicitly mention using web_fetch and curl to retrieve HTML and content from target websites for auditing crawlability, indexation, and on-page SEO (SKILL.md).
Boundary markers: The instructions do not provide explicit delimiters or "ignore instructions" to separate the untrusted fetched content from the agent's core task logic.
Capability inventory: The skill's intended behavior is to generate an "Audit Report Structure" which is natural language output. It does not contain instructions for system-level modifications, file writes, or arbitrary command execution.
Sanitization: The skill documentation notes that the web_fetch tool automatically strips <script> tags, which provides a level of protection against JavaScript-based injections, though it does not protect against instructions embedded in standard HTML or text.

seo-audit