The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because it ingests and processes untrusted data from external websites. Ingestion points: External web content retrieved via HarborRank tools like run_site_audit and get_audit_pages as described in SKILL.md. Boundary markers: Absent; the skill instructions do not provide delimiters or specific warnings for the agent to ignore instructions embedded within the crawled content. Capability inventory: The skill has the capability to write markdown files to the local filesystem (in the audits/ directory) and publish data to external URLs via the create_share_report tool (SKILL.md). Sanitization: Absent; no sanitization, escaping, or validation logic is specified for the external content before it is processed by the agent.

seo-audit