The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it instructs the agent to ingest and analyze external website content using tools like web_fetch and curl.
Ingestion points: The audit process involves fetching HTML and content from user-provided URLs in SKILL.md.
Boundary markers: There are no explicit instructions or delimiters provided to ensure the agent ignores potential malicious instructions embedded within the fetched website data.
Capability inventory: The skill is primarily designed for generating reports and recommendations, which limits the potential impact of an injection compared to skills with write access to the filesystem or network.
Sanitization: No sanitization or filtering of the fetched content is specified before the AI processes it.
[NO_CODE]: The skill contains no executable scripts (Python, Node.js, or Shell). It consists entirely of Markdown instructions, reference documentation, and JSON evaluation data.
[SAFE]: The skill demonstrates safe practices by attempting to read local business context from .agents/product-marketing-context.md and referencing official Google diagnostic tools for verification.

seo-audit