Skip to content

Security: lightfastai/lightfast

Security

SECURITY.md

Security Policy

Supported Versions

We actively maintain and provide security updates for the following versions:

Version Supported
Latest ✅
< Latest ❌

As Lightfast is in active development, use the latest published versions of the public CLI, SDK, and MCP packages for current security fixes.

Reporting a Vulnerability

We take security vulnerabilities seriously. Do not create a public GitHub issue for a suspected vulnerability.

Email security@lightfast.ai with:

  • A description of the vulnerability
  • Steps to reproduce it
  • Its likely impact
  • Any proposed fixes or mitigations
  • Contact information for follow-up

You can expect:

  • Acknowledgment within 48 hours
  • An initial assessment within five business days
  • Progress updates during the investigation
  • Coordinated disclosure for confirmed issues
  • Credit in security advisories, with your permission

Target response times are one to three business days for critical issues, one week for high-severity issues, and the next regular release cycle for medium- or low-severity issues.

Repository Security Boundaries

  • Public clients require explicit endpoint configuration; this repository does not provide an implicit hosted Lightfast backend.
  • apps/desktop is a sandboxed static local shell with no Node.js renderer integration, remote navigation, updates, signing, or distribution wiring.
  • apps/mcp is a local stdio-only shell with no HTTP listener, authentication, secret, deployment, or hosted endpoint.
  • Database clients require explicit credentials and do not connect at import time. Provider access and schema writes require separate approval.
  • Never commit credentials or copy package-local environment values between workspaces.

Dependency review, CodeQL, repository checks, and focused package tests run in CI. These controls supplement, rather than replace, review of trust boundaries and user-controlled inputs.

Security Best Practices

For Contributors

  • Validate user-controlled inputs at trust boundaries.
  • Never expose sensitive data in client-side code or error messages.
  • Keep dependencies current and review security advisories.
  • Use HTTPS for configured remote endpoints.
  • Preserve explicit endpoint configuration; do not add implicit hosted Lightfast defaults.
  • Keep provider access and environment values package-local.

For Users

  • Keep public Lightfast packages up to date.
  • Secure API keys and other environment variables.
  • Apply appropriate access controls at the configured compatible backend.
  • Monitor applications for unusual activity and back up important configuration.

Vulnerability Management

Repository maintainers use dependency review, automated security scanning, security-focused code review, and focused architectural review. Critical components may also receive targeted audits or penetration testing.

Contact

  • Security issues: security@lightfast.ai
  • General security questions: open a GitHub discussion with the security label, without including vulnerability details

Acknowledgments

We appreciate the security research community and acknowledge researchers who responsibly disclose vulnerabilities, with their permission.


This policy evolves with the project. Check back for current guidance.

There aren't any published security advisories