Kagi stores encrypted secrets. Security issues are taken seriously — please disclose responsibly.

Kagi is a self-hosted application distributed as source code. Only the latest commit on master receives security fixes. If you are running an older version, please update before reporting.

For context when evaluating reports:

• All secret values are encrypted with AES-256-GCM before being written to the database.
• Ciphertext format: iv:authTag:ciphertext (all base64, colon-separated).
• The master key (KAGI_ENCRYPTION_KEY) is supplied via environment variable and never stored in the database.
• Secret values are never returned by list or detail API endpoints — only via explicit POST /api/entries/[id]/reveal or POST /api/2fa/[id]/reveal, which require an authenticated session and ownership check.
• AI extraction is privacy-preserving: the model receives only key names and project names, never plaintext values.

Please do not open a public GitHub issue for security vulnerabilities.

Report via one of:

1. GitHub private security advisory (preferred): https://github.com/YanceyOfficial/kagi/security/advisories/new
2. Email: yanceyofficial@gmail.com — use the subject line [kagi] Security Vulnerability.

Please include:

• A description of the vulnerability and its potential impact
• Steps to reproduce or a proof-of-concept
• The component affected (auth, encryption, API route, etc.)
• Any suggested mitigations if you have them

You can expect an acknowledgement within 48 hours and a status update within 7 days. We will coordinate a fix and disclosure timeline with you.

Items that are in scope:

• Authentication/authorization bypass
• Encryption weaknesses or key exposure
• API routes that leak encryptedValue or plaintext secrets
• Session fixation or privilege escalation
• Injection vulnerabilities (SQL, command, etc.)

Items that are out of scope:

• Vulnerabilities requiring physical access to the server
• Misconfiguration of the operator's Keycloak or PostgreSQL instance
• Vulnerabilities in upstream dependencies (report those to the respective project)
• Self-XSS or issues requiring the attacker to have admin access