gladia-audio-intelligence

Pass

Audited by Gen Agent Trust Hub on Oct 22, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill enables features that process untrusted data (audio transcripts) which can be used to influence the agent's behavior.
  • Ingestion points: Transcription results from preRecorded().transcribe() (SKILL.md) and liveV2().startSession() (references/live-audio-intelligence.md) serve as the input for intelligence features.
  • Boundary markers: The code examples do not include explicit delimiters or instructions to ignore embedded commands within the transcripts.
  • Capability inventory: Transcripts are passed to Gladia's backend for summarization, chapterization, and 'Audio-to-LLM' custom prompt execution (references/pre-recorded-audio-intelligence.md).
  • Sanitization: No sanitization or validation of the generated transcript content is shown before it is processed by secondary AI features.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 22, 2026, 12:17 PM