Skip to content

Security: Compliledger/canton_tokenproof

Security

SECURITY.md

Security Policy

Supported versions

Version Supported
0.1.x ✅ Active

Reporting a vulnerability

Do not open a public GitHub issue for security vulnerabilities.

Report security issues by emailing the maintainers directly. Include:

  • A description of the vulnerability
  • The layer affected (DAML contracts / backend / SDK)
  • Steps to reproduce
  • Potential impact

You will receive an acknowledgement within 48 hours. We aim to release a fix or mitigation within 14 days for critical issues.


Scope

In scope

  • DAML authorization bypass — any path that allows an unauthorized party to create, archive, or exercise a choice on ComplianceProof, ComplianceGuard, or EvaluationRequest.
  • Backend authentication bypass — any path that allows unauthenticated access to ledger command submission in a non-sandbox environment.
  • Classification engine determinism violations — any input that causes engine.classify to return a different result for identical inputs on the same policy version.
  • Proof hash collision — any practical collision in the SHA-256 proof hash construction.

Out of scope

  • Canton sandbox running without authentication (this is the intended local development configuration).
  • submitMustFail test scripts exercising intentional failure paths.
  • SDK client-side issues that require a compromised backend.

Security model

TokenProof is a PoC / Canton Dev Fund milestone project. It is not yet hardened for MainNet production deployment. Specifically:

  • The current Canton adapter accepts an empty JWT for local sandbox use. In a production deployment, a valid JWT from a Canton participant node must be configured via CANTON_EVALUATOR_JWT.
  • The classification engine does not authenticate callers. In production, the /evaluate endpoint must be placed behind an authenticated API gateway.
  • The TOKENPROOF_PACKAGE_ID must be pinned to the deployed DAR hash. Mismatched package IDs will cause template resolution failures.

Disclosure policy

We follow coordinated disclosure. Once a fix is released, the vulnerability may be disclosed publicly with credit to the reporter.

There aren't any published security advisories