backend-websocket

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS] (SAFE): The skill installs standard, reputable npm packages (koa-websocket, short-uuid). No suspicious remote scripts or unverifiable binaries are downloaded.
  • [COMMAND_EXECUTION] (SAFE): The skill contains application-level TypeScript code. It does not invoke system shells, execute arbitrary commands, or utilize dangerous subprocess calls.
  • [CREDENTIALS_UNSAFE] (SAFE): No hardcoded secrets, API keys, or private tokens were found. The authentication logic uses a placeholder function, correctly prompting the developer to implement their own secure verification logic.
  • [DATA_EXFILTRATION] (SAFE): The skill does not access sensitive local file paths (such as ~/.ssh or .env) and only communicates via the defined WebSocket protocol for its intended purpose.
  • [PROMPT_INJECTION] (SAFE): The instructions are focused on implementation and do not contain any directives meant to override AI safety filters or extract system prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:22 PM