aiox-data-engineer

Warn

Audited by Gen Agent Trust Hub on Mar 16, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The activation protocol automatically executes 'node .aiox-core/development/scripts/generate-greeting.js data-engineer' when the skill is loaded. This pattern executes scripts from the local development environment without explicit user confirmation during the initialization phase.
  • [REMOTE_CODE_EXECUTION]: The skill relies on external script files and agent definitions located in '.aiox-core/' and '.codex/' that are not part of the skill's source code, creating a dependency on external code that is executed at runtime.
  • [PROMPT_INJECTION]: The '*yolo' command allows the user to cycle the agent's permission mode to 'auto' or 'explore'. These modes are designed to bypass human-in-the-loop safety constraints, which could result in the agent performing sensitive or destructive database operations, such as schema modifications or RLS policy updates, without manual approval.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 16, 2026, 11:44 PM