The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill requires the agent to ingest and analyze untrusted content from external websites using web_fetch and curl. This creates a potential surface for indirect prompt injection, where an attacker could embed malicious instructions in a website's HTML or metadata to influence the agent's behavior during the audit.
Ingestion points: External website content and metadata retrieved at runtime in SKILL.md.
Boundary markers: The instructions do not specify explicit delimiters or "ignore" directives for the fetched content.
Capability inventory: The skill facilitates web reading and local file access (.agents/product-marketing-context.md), but does not include capabilities for writing files or executing shell commands based on the external input.
Sanitization: The skill lacks explicit instructions for sanitizing or filtering the content retrieved from audited URLs.

seo-audit