The Agent Skills Directory

[SAFE]: The skill is designed for local identity synthesis and interview orchestration. It does not utilize any network tools or external data fetching mechanisms.
[SAFE]: The instructions incorporate proactive security measures by mandating the validation of participant names (alphanumeric only) and outputPath (absolute path verification, rejection of .. segments). These constraints effectively mitigate potential path traversal vulnerabilities.
[SAFE]: The use of disk-persisted state (state.json) is limited to a predefined project directory, and the skill includes logic to ensure that session data is handled within safe boundaries.

brand-discovery