We release patches for security vulnerabilities for the following versions:

Please do not report security vulnerabilities through public GitHub issues.

Instead, please report them via email to: security@corebasehq.com

You should receive a response within 48 hours. If for some reason you do not, please follow up via email to ensure we received your original message.

Please include the following information:

• Type of vulnerability
• Full paths of source file(s) related to the vulnerability
• Location of the affected source code (tag/branch/commit or direct URL)
• Any special configuration required to reproduce the issue
• Step-by-step instructions to reproduce the issue
• Proof-of-concept or exploit code (if possible)
• Impact of the issue, including how an attacker might exploit it

This information will help us triage your report more quickly.

When we receive a security bug report, we will:

1. Confirm the problem and determine affected versions
2. Audit code to find any similar problems
3. Prepare fixes for all supported releases
4. Release patched versions as soon as possible

When using CoreMCP:

1. Never commit sensitive credentials to version control

   • Use .gitignore to exclude coremcp.yaml
   • Use environment variables for sensitive data

2. Use read-only database accounts whenever possible

   sources:
     - name: "production_db"
       readonly: true  # Enforce read-only mode

3. Restrict database permissions

   • Grant only SELECT permissions to the database user
   • Use dedicated service accounts with minimal privileges

4. Secure connection strings

   • Use encrypted connections when possible
   • For MSSQL: Add encrypt=true to DSN
   • Store DSN strings in secure configuration management

1. Use stdio transport for local Claude Desktop integration

   • No network exposure
   • Recommended for most use cases

2. If using HTTP transport (future):

   • Use HTTPS only
   • Implement authentication
   • Use firewall rules to restrict access

1. Keep CoreMCP updated to the latest version
2. Monitor logs for suspicious activity
3. Audit database access regularly
4. Use principle of least privilege

server:
  name: "coremcp-agent"
  version: "0.1.0"
  transport: "stdio"  # Secure local transport

logging:
  level: "info"
  format: "json"

sources:
  - name: "production_db"
    type: "mssql"
    # Use environment variable: ${COREMCP_SOURCES_0_DSN}
    dsn: "sqlserver://readonly_user:${DB_PASSWORD}@localhost:1433?database=mydb&encrypt=true"
    readonly: true  # Enforce read-only

Set sensitive values via environment:

export COREMCP_SOURCES_0_DSN="sqlserver://user:password@host:1433?database=db&encrypt=true"
./coremcp serve

• CoreMCP passes queries directly to the database
• The AI assistant constructs SQL queries
• Use read-only database accounts to limit potential damage
• Consider implementing query allowlists for production use

• Query results are returned to the AI assistant
• Be aware of what data is accessible
• Use database views to limit data exposure
• Apply row-level security in your database

Security updates will be released as patch versions (e.g., 0.1.1, 0.1.2) and announced via:

• GitHub Security Advisories
• Release notes
• Email to registered users (if applicable)

If you have suggestions on how this policy could be improved, please submit a pull request or open an issue.

Last updated: February 14, 2026