Skip to content

Security: rushillshah/digital-brain

Security

SECURITY.md

Security Policy

Digital Brain handles private local data, app databases, generated memory, API keys, browser sessions, and optional outbound messaging. Treat security and privacy reports as high priority.

Supported Versions

The latest npm release and main branch receive security fixes.

Reporting A Concern

Do not open public issues that include secrets, private messages, database paths, auth tokens, QR/session data, or exploit details.

Open a minimal public issue titled Security: private report needed, or contact the maintainer privately if you already have a private channel.

Defaults

  • No cloud API is called by default.
  • Messages and generated memory are stored as local files.
  • Telemetry is opt-in, anonymous, and content-free.
  • Outbound sending requires explicit configuration and should be tested in draft mode first.
  • ZIP importers reject archive entries that try to escape the extraction directory.
  • Local config files are written with owner-only permissions on POSIX systems where supported.
  • Chromium sandboxing stays enabled by default. Set DIGITAL_BRAIN_CHROME_NO_SANDBOX=1 only in constrained environments that require it.

High-Risk Areas

  • WhatsApp Web sessions and send automation.
  • Local WhatsApp and iMessage databases.
  • Provider prompts sent to OpenAI, Anthropic, xAI, Codex CLI, or other configured providers.
  • Logs, telemetry, screenshots, and generated vault files.
  • Raw source data under 08 Sources/.

Maintainer Checklist

  • Never ask users to paste real chats into public issues.
  • Never commit .npmrc, API keys, browser sessions, generated vaults, or local databases.
  • Keep raw data, derived data, and AI-readable memory boundaries visible.
  • Make network/API behavior explicit in docs and prompts.

There aren't any published security advisories