The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted content from external websites, creating a vulnerability surface for indirect prompt injection.\n- Ingestion points: Untrusted data enters the context via web_fetch and curl as outlined in the Audit Framework and Technical SEO sections of SKILL.md.\n- Boundary markers: The prompt does not utilize delimiters or specific instructions to treat external content as data rather than instructions.\n- Capability inventory: The skill performs network operations and suggests the use of browser tools to execute JavaScript snippets for schema validation.\n- Sanitization: The skill acknowledges that web_fetch strips script tags, but it lacks mechanisms to sanitize textual content to prevent instruction override.

seo-audit