telnyx-voice-gather-ruby

Pass

Audited by Gen Agent Trust Hub on Mar 27, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill uses the official 'telnyx' Ruby gem to interact with Telnyx APIs. All network operations are directed at the service's own infrastructure.
  • [CREDENTIALS_UNSAFE]: The skill correctly recommends using environment variables for managing sensitive API keys instead of hardcoding them.
  • [DATA_EXFILTRATION]: No unauthorized data exfiltration patterns were detected. Data processed by the skill is limited to call control parameters and interaction results sent to or received from the Telnyx service.
  • [COMMAND_EXECUTION]: No suspicious shell command execution was found. The only shell command mentioned is the standard installation of the official SDK via 'gem install'.
  • [PROMPT_INJECTION]: The skill processes untrusted data from external sources, specifically webhook payloads containing caller input (DTMF digits and speech transcripts). This constitutes a surface for Indirect Prompt Injection.
  • Ingestion points: The skill reads the request body in the webhook handler and processes fields like 'data.payload.digits' and 'data.payload.result'.
  • Boundary markers: Not explicitly shown in the code snippets.
  • Capability inventory: The skill can send messages and initiate call actions via the client object.
  • Sanitization: The skill explicitly demonstrates and recommends webhook signature verification using Ed25519 signatures, which ensures the authenticity of the data source.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 27, 2026, 01:02 AM