Please do not open a public issue for security problems.
Report privately via GitHub's Report a vulnerability button (the repo's Security tab, then Advisories). If you can't use that, email security@openbower.ai instead.
We aim to acknowledge within 3 business days and will keep you updated as we investigate. Once a fix ships we're glad to credit you, unless you'd rather stay anonymous.
OpenBower is self-hosted and BYO-key: AI models and search providers run on credentials you supply, against your own instance. The areas we care most about:
- Authentication and account scoping (sessions, the OAuth client).
- The agent research path: fetching and searching against untrusted URLs (SSRF containment) and prompt-injection handling of fetched content.
- Handling of user-supplied provider API keys and any secrets at rest.
- The CSV import parser (untrusted file input).
Pre-1.0: fixes land on main. Pin a commit if you need stability.