retellai-data-handling

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to handle and process call transcripts which contain untrusted text from external callers. 1. Ingestion points: The redactTranscript and webhook patterns in references/data-patterns.md ingest transcript utterances. 2. Boundary markers: The skill does not define specific delimiters or instructions to prevent the agent from interpreting transcript text as instructions. 3. Capability inventory: The skill utilizes Write, Edit, and Bash tools, creating a risk surface if the agent is manipulated by transcript content. 4. Sanitization: The skill provides a redactText function using regex to filter PII like phone numbers and emails, though it does not specifically sanitize for prompt injection.
  • [EXTERNAL_DOWNLOADS]: The skill references the retell-sdk Node.js package. This is the official SDK for the Retell AI platform and is considered a safe and expected dependency for this integration.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:46 AM
Security Audit — agent-trust-hub — retellai-data-handling