The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied queries and domain names and retrieves responses from external LLM APIs, creating a vulnerability surface where instructions embedded in external data could influence agent behavior.
Ingestion points: User-provided keyword, domain, and query parameters defined in the command reference of SKILL.md.
Boundary markers: Absent; user input is interpolated directly into the JSON body of curl requests.
Capability inventory: Shell command execution via Bash and network operations using curl to external endpoints.
Sanitization: No input validation or response filtering mechanisms are described in the skill instructions.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill accesses a sensitive local file path at ~/.config/dataforseo/auth to retrieve authentication credentials.
The credentials are used for legitimate communication with the official api.dataforseo.com domain, which is a well-known SEO service provider.

30x-seo-ai-visibility