Skip to content

Security: ColdAI-org/clprouter

Security

SECURITY.md

Security policy

Reporting a vulnerability

Report vulnerabilities privately. Do not open a public issue, pull request or discussion.

  • E-mail: security@REPLACE-WITH-PROVIDER-DOMAIN (placeholder: the maintainers set the real address before the first public release)
  • Or use GitHub's private vulnerability reporting on this repository ("Security" tab, "Report a vulnerability").

Please include: the affected file and commit, a description, the impact you expect, and a proof of concept (a Foundry test is ideal). Encrypt sensitive details if you can; the maintainers publish a PGP key alongside the address.

We aim to acknowledge within 3 working days, give a first assessment within 10 working days, and agree a disclosure date with you. We credit reporters who want credit.

Scope

In scope Out of scope
src/ (Router, registry, vault, libraries, interfaces) lib/clpr-smart-contracts: report to the CLPR project (LFDT-CLPR) under its own policy
proto/ and the codecs in src/libraries/RouteCodec.sol and sdk/src/ Test helpers, scripts and fixtures
sdk/ (planner, envelope builder, ISO 20022 module) Third-party dependencies, unless CLPRouter uses them unsafely
services/ and the release container image Issues already listed in docs/threat-model.md section 7, unless you show a worse impact
CI and release workflows in .github/ Social engineering, physical attacks, denial of service by volume

Supported versions

Version Supported
0.1.x Yes (local networks and testnets only; not deployed on any mainnet)

How fixes ship

The contracts are immutable. A fix to an on-chain bug ships as a new Router version deployed beside the old one; the provider committee can disable an unsafe version (docs/deployment.md, section 5). Off-chain fixes ship as new SDK and services releases.

There aren't any published security advisories