English | 中文

The Sakura AI team takes the security of our application, infrastructure, and user data seriously. We appreciate the responsible disclosure of vulnerabilities by the security community and our users.

Security updates and patches are actively maintained for the following versions and branches:

If you are running an older unsupported version, please upgrade to the latest stable release before filing a vulnerability report.

Please do NOT report security vulnerabilities through public GitHub Issues, Pull Requests, or public discussion channels.

We strongly recommend reporting vulnerabilities using GitHub's built-in private reporting feature:

1. Navigate to the repository's Security tab: https://github.com/Sakura520222/Sakura-AI/security
2. Click "Advisories" in the left sidebar.
3. Click "Report a vulnerability" (or use the direct link: New Security Advisory).
4. Provide a detailed summary, affected versions, severity assessment, and reproduction steps.

This creates a private workspace where you can collaborate securely with the repository maintainers.

If you cannot use GitHub Security Advisories or need to attach sensitive materials, please send an encrypted or direct email to the project maintainer:

• Contact Email: sakura520222@outlook.com
• Subject Line: [SECURITY VULNERABILITY] Sakura AI - <Brief Description>

To help us triage and resolve the issue quickly, please include as much information as possible:

• Vulnerability Category: (e.g., Remote Code Execution, Authentication Bypass, Injection, SSRF, Information Disclosure, Sandbox Escape).
• Affected Component: Specific service, endpoint, background worker, or updater module.
• Affected Versions / Commit Hash: The environment where the issue was discovered.
• Proof of Concept (PoC): Clear, step-by-step reproduction instructions or exploit scripts.
• Impact Assessment: The potential impact if exploited by an attacker.
• Suggested Fix: Any recommended patches or mitigations (optional).

1. Initial Acknowledgment: We aim to acknowledge receipt of your vulnerability report within 48 hours.
2. Investigation & Triage: We will verify the vulnerability, evaluate its severity (CVSS score), and keep you informed of our progress.
3. Patch Development & Testing: A fix will be developed in a private security advisory branch and thoroughly tested.
4. Coordinated Disclosure:
   • Once a patched release is published, we will publish the GitHub Security Advisory.
   • We follow responsible disclosure principles and request a 90-day embargo period from the initial report to give users adequate time to patch their installations.
   • You will be properly credited in the security advisory release notes (unless you prefer anonymity).

If you are self-hosting Sakura AI, please follow these baseline recommendations:

• Disable Debug Mode in Production: Ensure DEV_BOOTSTRAP and debug flags are turned off.
• Protect Credentials: Keep config/connection.json and .deploy/deployment.env restricted with strict OS file permissions (chmod 600).
• Rotate Secrets: Regularly rotate GitHub App private keys, webhook secrets, database credentials, and session tokens.
• Enable MFA: Enforce Two-Factor Authentication (TOTP / Passkeys) in the WebUI Security Center for all administrative accounts.
• Network Isolation: Do not expose internal services (MySQL, Redis, ChromaDB, Updater unix socket) directly to the public internet; use firewalls, Docker networks, or Cloudflare Tunnels.