Detect Intents

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it processes untrusted meeting transcripts and translates them into automated system actions.
  • Ingestion points: The skill consumes a 'transcript' input string which originates from external meeting recordings (SKILL.md).
  • Boundary markers: The agent instructions do not provide explicit delimiters or clear directives to ignore instructions that might be embedded in the transcript text.
  • Capability inventory: The skill can trigger various platform tools, including those for proposal generation, calendar management, and system access provisioning.
  • Sanitization: There is no evidence of transcript content validation or sanitization to prevent adversarial manipulation of the intent detection process.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 11:37 AM