Brainstorm's core promise is that your data stays yours: apps are sandboxed, every privileged call crosses a capability broker that fails closed, sync is end-to-end encrypted, and keys live in the OS keychain. A hole in any of that is the most important bug we can hear about — thank you for taking the time to report it.

Please do not open a public issue for anything security-sensitive.

• Preferred: Report privately via GitHub (Security tab → Report a vulnerability). This keeps the report, the discussion, and any fix coordinated in one private place.
• Alternative: email founder@getbrainstorm.online.

Include what you can: the Brainstorm version (Settings → What's new), platform, steps to reproduce, and your assessment of impact. A proof of concept helps enormously; please demonstrate it against your own vault and data only.

You can expect an acknowledgment within 72 hours and an assessment within a week. Brainstorm is maintained by a very small team, so we ask for reasonable time to ship a fix before public disclosure — we'll agree on a timeline with you in the report thread, and we'll credit you in the release notes unless you'd rather stay anonymous. There is currently no bug bounty program.

Reports of highest interest — the boundaries the product's security model stands on:

• App-sandbox escape — a sandboxed app reading or writing anything it wasn't granted, reaching Node/Electron APIs, or escaping its renderer.
• Capability-broker bypass — any path to a host service that skips or defeats the IPC broker, the identity stamping, or the per-vault capability ledger; anything that makes a denied request succeed.
• Cross-app isolation failures — one app observing or influencing another app's data or execution.
• Cryptography flaws — in vault encryption, end-to-end encrypted sync, device pairing, collection sharing / key rotation, or identity signatures.
• Credential exposure — API keys, sync keys, or keychain-held secrets reaching an app, a log, a crash report, or the network.
• Update-mechanism attacks — anything that could make the in-app updater install an unintended artifact.
• AI-boundary failures — model-directed actions escaping their granted capabilities or spend budgets, including via prompt injection from vault content.

Out of scope: issues requiring full control of the user's OS account or physical machine, denial of service against your own local instance, vulnerabilities purely in third-party dependencies with no Brainstorm-specific exploit path (report upstream — though a heads-up is welcome), and social engineering.

Brainstorm is in public beta. Security fixes ship in new releases on the latest release line — always via the current release and the in-app updater (Settings → Updates). Older builds are not patched retroactively.