diabatix-coldstream

Pass

Audited by Gen Agent Trust Hub on May 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the @membranehq/cli package globally via NPM. This is a vendor-provided tool required for the skill to interact with the Membrane platform.
  • [COMMAND_EXECUTION]: The skill relies on executing shell commands through the membrane CLI to manage connections, discover capabilities, and perform actions within the Diabatix ColdStream environment.
  • [DATA_EXPOSURE]: The skill follows secure secret management practices by explicitly instructing the agent never to ask users for API keys or tokens, instead utilizing Membrane's server-side connection management.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from external actions (results of membrane action run) and lists (output of membrane action list). While this presents a potential attack surface for indirect prompt injection from compromised thermal design data or action descriptions, the risk is inherent to the integration's purpose and mitigated by standard LLM guardrails.
Audit Metadata
Risk Level
SAFE
Analyzed
May 3, 2026, 09:21 PM
Security Audit — agent-trust-hub — diabatix-coldstream