The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided parameters to configure scheduled network requests, establishing a surface for potential indirect prompt injection attacks. * Ingestion points: Arguments including url, headers, and body in the create_schedule tool call (SKILL.md). * Boundary markers: No delimiters or instructions to ignore embedded commands are specified for the user-supplied input. * Capability inventory: The skill facilitates network operations (HTTPS requests) through an external scheduling service. * Sanitization: The instructions do not include specific sanitization, filtering, or validation steps for the external data.
[DATA_EXFLITRATION]: The skill facilitates network operations to external service endpoints as part of its core functionality. * Evidence: SKILL.md references network interactions with api.crontap.com and supabase.co. These operations target the vendor's own infrastructure and a well-known service provider, respectively.

crontap-supabase-edge-function-cron