Flag for Manager
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
execute_actionfunction to interface with CRM and task management systems. This is used for standard operations such as fetching deal health signals and creating follow-up tasks. - [DATA_EXFILTRATION]: Deal-related data and risk assessments are sent to Slack. This behavior is restricted to internal communication directed toward the user's manager as defined in the platform's user profile.
- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it ingests data from external sources like CRM activity logs and meeting transcripts. While malicious instructions in these sources could potentially influence the AI's risk assessment or summary text, the risk is considered low as the agent's output is limited to a Slack notification and a tracking task.
Audit Metadata