Meeting Prep Brief

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill logic and instructions contain no evidence of malicious intent, obfuscation, or unauthorized data exfiltration. It operates using standard platform capabilities for data retrieval.
  • [PROMPT_INJECTION]: The skill identifies an attack surface for indirect prompt injection by processing untrusted data from external sources.
  • Ingestion points: Meeting titles, descriptions, and transcripts are retrieved via get_meetings and transcript search actions.
  • Boundary markers: The prompt does not define explicit delimiters to isolate untrusted text from instructions.
  • Capability inventory: The skill has no access to critical execution tools such as shell commands, file writing, or network operations, significantly limiting the impact of potential injection.
  • Sanitization: No explicit sanitization or filtering of external text content is implemented.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 11:37 AM