scheduling
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides standard instructional content for scheduling tasks. It does not contain any malicious code, obfuscated commands, or unauthorized data access patterns.
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, creating a potential attack surface. However, the instructions incorporate significant defensive behaviors to mitigate these risks:
- Ingestion points: Reads data from calendars, external booking channels, and email communications via
$orbit:email(SKILL.md). - Boundary markers: Explicitly references
<confirmation_policy>and$orbit:writing-styleto constrain output and behavior (SKILL.md). - Capability inventory: Capabilities include reading/writing calendar events and sending emails (SKILL.md).
- Sanitization: The instructions require the agent to 'Verify the saved date, time zone, attendees, place, meeting link and recurrence' and 'recheck availability before booking' to ensure data integrity (SKILL.md).
Audit Metadata