MemoryAI takes the security and privacy of AI agent memory extremely seriously.

If you discover a potential security vulnerability in MemoryAI, please report it immediately by emailing security@memoryai.org or submitting a confidential security advisory via GitHub.

Please do NOT open public GitHub issues for security vulnerabilities.

1. Description of the vulnerability (e.g. IDOR, SSRF, prompt injection bypass, authentication flaw).
2. Step-by-step reproduction steps or Proof-of-Concept script.
3. Impact assessment and suggested remediation.

• Acknowledgment: Within 24 hours.
• Initial Assessment: Within 72 hours.
• Remediation & Patch Release: Within 7 days for critical vulnerabilities.

1. OWASP API Security Top 10 Hardened: Strict Tenant, User, and Project authorization checks on every operation to prevent Insecure Direct Object References (IDOR).
2. SSRF Guard: Strict DNS verification blocking loopback (127.0.0.1), private networks (RFC1918), link-local addresses, and cloud instance metadata endpoints (169.254.169.254).
3. Prompt Injection Isolation: Retrieved memory data is strictly framed inside <MEMORY_DATA> containers with explicit non-executable attribution headers.
4. Encryption at Rest: Sensitive fields are encrypted using authenticated AES-256-GCM.
5. Non-Root Containers: Docker containers drop all unnecessary Linux capabilities and execute as an unprivileged non-root user (UID 10001).