Skip to content

Security: GlassAlpha/glassalpha

SECURITY.md

Security policy

Do not create public issues for security vulnerabilities, instead contact us at [email protected]

Our security commitment

GlassAlpha is designed for regulated industries requiring high trust and transparency. Security is fundamental to our mission of providing audit-ready ML compliance tools.

Reporting security vulnerabilities

If you discover a security vulnerability in GlassAlpha, please help us protect our users by reporting it responsibly:

How to report

  • Email: [email protected]
  • Response Time: We aim to acknowledge reports within 48 hours
  • Process: We will investigate all reports and provide updates on our progress

What to include

  • Description of the vulnerability and its potential impact
  • Steps to reproduce the issue
  • Your assessment of severity
  • Any suggested fixes or mitigations

What NOT to do

  • Do not create public GitHub issues for security vulnerabilities
  • Do not publicly disclose the vulnerability before we've had a chance to address it

Supported versions

Version Supported
0.1.x ✅ (Pre-alpha development)

Security design principles

GlassAlpha follows these security principles:

  • On-Premise First: No external network calls or cloud dependencies by design
  • Privacy by Default: No telemetry collection (opt-in only via GLASSALPHA_TELEMETRY=on)
  • Data Protection: Never logs raw PII; all identifiers are hashed
  • Reproducible Security: All operations are deterministic and auditable

Security features

  • Local Processing: All model analysis happens on your infrastructure
  • No Data Transmission: Your data never leaves your environment
  • Audit Trail: Complete lineage tracking with cryptographic hashes
  • Deterministic Output: Identical results for compliance verification

Reporting security vulnerabilities

Please report security vulnerabilities to us at [email protected]

Questions?

Contact us: [email protected]

There aren’t any published security advisories