Bob Labs runs as infrastructure on your own servers — most deployments handle private RAG, credentials for LLM providers, GPU agents, and (optionally) Web3 keys. A vulnerability here can have outsized impact, so we take reports seriously and ask you to disclose them privately first.

Please do NOT open a public GitHub issue for security-relevant findings.

Email support@boblabs.eu with:

• A short description of the issue and its impact.
• Steps to reproduce (or a proof-of-concept, if you have one).
• The affected component(s) — bob-api, bob-ui, bob-agent, sandbox, gpu-services, etc.
• The commit SHA or release tag you tested against.

We aim to acknowledge every report within 7 days, share an initial severity assessment within 14 days, and ship a fix or mitigation as quickly as the issue warrants. If you'd like credit in the fix's commit message or release notes, say so in the report.

In scope:

• Remote code execution, sandbox escape, or unauthorized container access.
• Authentication or authorization bypass (operator JWT, consumer-app HMAC, access tokens).
• Cross-tenant data leakage (RAG collections, lab workspaces, agent memories).
• Secret exposure through logs, error responses, or admin UIs.
• SQL injection, SSRF, or path traversal against any HTTP endpoint.
• LLM prompt-injection paths that reach a privileged tool (e.g. shell_exec, trading, mail.send).

Out of scope:

• Findings that require operator-level access you already have (the operator is trusted; that is the trust model).
• Best-practice suggestions ("you should rotate JWTs every 7 days") without a concrete exploit.
• Denial-of-service via raw resource exhaustion — labs are CPU/GPU-bound by design; the right knobs are rate limits and quotas in your reverse proxy / orchestrator.
• End-of-life version compromises — please test against main or the latest release tag.

We follow standard coordinated disclosure: we ask reporters to wait until a fix has shipped (or 90 days have passed, whichever comes first) before publishing details. We will work with you to align on timing if the issue is being actively exploited.

Thanks for helping keep Bob Labs deployments safe.