The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill enables the agent to process inbound message data from webhooks, which includes user-controlled text fields. This creates a vector where malicious instructions could be embedded in messages to influence agent behavior.\n
Ingestion points: Inbound message payloads containing user-supplied text and referral fields as described in SKILL.md.\n
Boundary markers: The documentation does not provide specific instructions for the use of delimiters or warnings to ignore instructions within the processed data.\n
Capability inventory: The skill provides methods for sending messages and interacting with the Zavu API, which could be abused if the agent is manipulated by untrusted input.\n
Sanitization: There is no mention of sanitizing or validating the incoming text content before it is processed by the agent.\n- [COMMAND_EXECUTION]: The skill documents the use of the zavudev CLI tool through npx for managing webhook settings. These commands are intended for administrative configuration of the sender's webhook parameters and represent intended vendor functionality.

webhook-setup