The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because its primary function involves ingesting and analyzing untrusted content from external websites. This could allow an attacker to embed malicious instructions on a webpage to influence the agent's behavior during an audit.
Ingestion points: The skill instructs the agent to use tools like web_fetch or curl to analyze external websites and to read workspace context from files such as .agents/product-marketing.md.
Boundary markers: There are no instructions to use boundary markers, delimiters, or explicit "ignore embedded instructions" warnings when processing data from these external sources.
Capability inventory: The agent is tasked with web fetching and generating diagnostic reports; however, the skill itself does not provide custom scripts with dangerous file-system or network write capabilities.
Sanitization: The instructions do not prescribe any specific sanitization or filtering of the content fetched from the web before the agent processes it.

seo-audit