Organization travel status to meeting attendees

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill implements logic for scheduling meetings based on organizational membership and travel status observations.
  • [DYNAMIC_EXECUTION]: The validator.py script uses standard validation utilities for schema checks and does not execute arbitrary commands or generated code.
  • [DATA_EXFILTRATION]: No network calls or sensitive file access patterns were found in any of the skill files.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes dialogue data to extract meeting parameters. While it lacks explicit boundary markers, it does not possess the high-risk capabilities (e.g., network access, shell execution) required for exploitation.
  • Ingestion points: Meeting details extracted from dialogue as described in rules.md.
  • Boundary markers: Absent; instructions do not specify delimiters for dialogue content.
  • Capability inventory: Limited to the AddMeeting tool call.
  • Sanitization: Time normalization is specified in the parameter rules.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 05:13 PM
Security Audit — agent-trust-hub — Organization travel status to meeting attendees