Skip to content

SECURITY.md

Latest commit

 

History

History
48 lines (34 loc) · 2.71 KB

SECURITY.md

File metadata and controls

48 lines (34 loc) · 2.71 KB
Raw

Security Policy

Reporting a Vulnerability

If you discover a security vulnerability in Veritas, please report it responsibly.

Email: Open a GitHub issue with the label security. If the vulnerability is sensitive (e.g., it could be exploited before a fix is deployed), please reach out privately before disclosing publicly.

Please include:

  • A description of the vulnerability
  • Steps to reproduce
  • The potential impact
  • Suggested fix (if you have one)

Security Model

Veritas is a cryptographic verification platform. Its security relies on:

  • Ed25519 signatures — All state-changing actions require valid cryptographic signatures. There is no anonymous mode.
  • Commit-reveal protocol — Results are sealed (hashed) before any are revealed, preventing validators from copying each other.
  • Hash-chained audit log — Every event is chained via SHA-256, making tampering detectable.
  • Verifiable random selection — Validator selection uses drand beacon randomness, so anyone can verify the draw was fair.
  • Parameterized SQL — All database queries use parameterized placeholders (no SQL injection surface).
  • Serializable transactions — Critical state transitions use SERIALIZABLE isolation to prevent race conditions.

Known Limitations

Veritas is pre-production software. The following are known areas to address before production deployment:

  • CORS is unrestricted — All origins are currently allowed. Production deployments should whitelist origins.
  • Rate limiter is IP-based — Behind a reverse proxy, all requests appear from the same IP. X-Forwarded-For support is not yet implemented.
  • Docker image runs as root — The production Dockerfile does not set a non-root USER.
  • Server keys are ephemeral by default — Without VERITAS_SERVER_KEY configured, a fresh Ed25519 keypair is generated on each startup.
  • No pagination on pool list endpoints — Large result sets could cause high memory usage.
  • Verification-specific threats — See docs/security-audit.md for the full audit, including threats specific to the cross-validation protocol (colluding validators, cache poisoning, replay attacks).

Cryptographic Dependencies

Algorithm Library Purpose
SHA-256 crypton Hashing, content addressing, audit log chaining
Ed25519 crypton Participant and server signatures
BLS12-381 hsblst drand beacon signature verification
HKDF-SHA256 crypton Key derivation

Full Audit

A detailed security audit with findings, threat analysis, and recommendations is maintained at docs/security-audit.md.