AgentOS is in v0.1.0-alpha. Security fixes are applied to the latest main branch.

If you discover a security vulnerability in AgentOS, please report it responsibly.

Do NOT open a public GitHub issue for security vulnerabilities.

Instead, please email: security@nousresearch.com

Include the following information:

1. Description of the vulnerability
2. Steps to reproduce
3. Potential impact
4. Suggested fix (if any)

You will receive a response within 72 hours. If the vulnerability is confirmed, we will work on a fix and coordinate disclosure with you.

AgentOS includes a multi-layer security model documented in:

• docs/adr/005-security-hypervisor.md — Security Hypervisor architecture
• docs/constitution/threat-model-v1.md — Threat model and mitigations
• docs/constitution/acp-v1.md — Agent Communication Protocol (Ed25519 signing, encryption)

Key security features:

• All agent communication is Ed25519 signed (no unsigned messages accepted)
• 6-layer Security Hypervisor with pre-invoke and post-invoke checks
• Permission engine with workspace-scoped grants
• Sandboxed capability execution
• Constitutional invariants enforced at runtime (10 invariants, 0 violations in benchmarks)

• We follow coordinated disclosure
• Credit will be given to reporters (unless they prefer to remain anonymous)
• We request a 90-day disclosure window before public publication