deal-tracker
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface because it processes untrusted data from external emails, Slack messages, and project management tools to generate summaries and automate communication. This is a common characteristic of CRM-style skills.
- Ingestion points: Reads external content through tools like Superhuman_Mail.get_thread and slack_search_public_and_private.
- Boundary markers: Absent. The instructions do not specify delimiters or protective prompts to isolate external content from the agent's instructions.
- Capability inventory: The skill can draft emails and create calendar events using Superhuman_Mail.create_or_update_draft and Superhuman_Mail.create_or_update_event.
- Sanitization: Absent. There are no explicit instructions for sanitizing or filtering retrieved content before it is processed by the agent.
Audit Metadata