The Agent Skills Directory

[SAFE]: No malicious patterns, encoded payloads, or unauthorized network operations were found in the clinical report guidelines or templates. The skill's focus is on professional documentation and regulatory compliance.
[PROMPT_INJECTION]: The skill handles untrusted external clinical data, presenting an indirect prompt injection surface. * Ingestion points: Clinical documentation and diagnostic reports processed by the agent. * Boundary markers: Instructions do not specify delimiters or warnings for the agent to disregard instructions within the clinical data. * Capability inventory: The skill is permitted to use Bash, Write, Edit, and Read tools. * Sanitization: There are no explicit requirements for sanitizing or escaping medical content before processing.
[COMMAND_EXECUTION]: The skill documentation references local Python scripts in the scripts/ directory for validation and compliance tasks, which are executed via the Bash tool. These tools appear to be standard, benign utilities for the intended clinical reporting workflow.

clinical-reports