terra-streaming

Pass

Audited by Gen Agent Trust Hub on Oct 29, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: Instructions in the documentation suggest installing the official Terra CLI by downloading a script from the vendor's domain and executing it directly in the shell.
  • Evidence: curl -fsSL https://cli.tryterra.co/install.sh | sh in SKILL.md.
  • Evidence: irm "https://cli.tryterra.co/install.ps1" | iex for Windows environments in SKILL.md.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the cloning of example projects and the downloading of configuration data from official vendor repositories.
  • Evidence: Fetching REST endpoints and protocol frames from docs.tryterra.co.
  • Evidence: Cloning project templates using terra examples clone.
  • [COMMAND_EXECUTION]: The skill utilizes a CLI tool to manage developer environments, list data tokens, and administer users.
  • Evidence: Usage of terra environments list, terra users list, and terra data-tokens create throughout SKILL.md.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to process real-time biometric telemetry (heart rate, ECG, location, etc.) received from external hardware sensors, which is an untrusted data source.
  • Ingestion points: Biometric data arrives at the backend via websocket DISPATCH payloads in JSON format (SKILL.md, references/consumer-protocol.md).
  • Boundary markers: No specific delimiters are defined as the broker passes sensor payloads opaquely.
  • Capability inventory: The skill provides the logic for backends to consume and react to these data streams.
  • Sanitization: Data is treated as raw telemetry; sanitization of this input for downstream use cases is the responsibility of the implementing developer.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 29, 2026, 12:15 AM