The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes data from the Kit API that could contain hidden instructions (indirect prompt injection) which may influence the agent's behavior.
Ingestion points: Untrusted data enters the agent context through tools like KIT_LIST_SUBSCRIBERS (subscriber metadata) and KIT_GET_BROADCAST (email content).
Boundary markers: The instructions do not specify any boundary markers or delimiters to help the agent distinguish between data and system instructions.
Capability inventory: The skill includes tools with destructive capabilities, such as KIT_DELETE_SUBSCRIBER and KIT_DELETE_BROADCAST.
Sanitization: There is no mention of sanitizing or validating the ingested content before it is processed by the agent.

convertkit-automation