The Agent Skills Directory

Local Configuration Access: The skill reads risk calibration data from its internal configuration directory. This enables the agent to provide consistent triage answers based on predefined organizational standards.
Privilege and Data Safeguards: A 'Destination check' is incorporated into the workflow to evaluate where the agent's response will be shared. This serves as a safety mechanism to help users preserve attorney-client privilege and avoid inadvertent disclosure in public environments.
Input Context Processing: The skill analyzes user-provided questions to match them against a calibration table. While the skill processes external text, it lacks network access or high-privilege command execution, which limits the potential impact of processing untrusted data.

is-this-a-problem