Security & Trust

OmeraCode is engineered with a multi-layered security architecture to ensure your firmware development workflow remains fully governed.

1. Project Workspace Sandbox

All AI analysis and code generation tasks are isolated within the active project workspace directory.

• Escape Prevention: Strict directory boundary validation prevents Path Traversal or Symlink Escape exploits.
• Command Risk Allowlist: Only explicit, pre-authorized CLI commands are permitted (`doctor`, `fw build run`, `fw build explain`, `fw debug`). Unregistered commands are blocked at execution choke points.

2. Device Authentication & Secrets

We utilize standard OAuth device authorization flow (similar to GitHub CLI) to authenticate sessions securely without requiring username/password input in the terminal. Tokens are saved locally inside your OS SecretStorage.

3. Data Redaction Gate

Underlying AI engines are completely redacted and unified under Omera 3.6 across all system logs, session histories, reports, and telemetry. Credentials and environment configurations are never cached or transmitted.

4. Vulnerability Disclosure

If you discover any security issues within our tooling, please report them directly to: support@omeraeda.com. We acknowledge and address reports within 48 hours.