Invarcore develops zero-trust, governance-critical, and verification-focused infrastructure for autonomous AI systems. We treat all security vulnerabilities with the highest priority.
We provide security updates and patches for the following versions:
| Repository / Engine | Active Version | Security Support |
|---|---|---|
canary-fabric |
v0.2.x |
✅ Supported |
intent-fabric |
v0.1.x |
✅ Supported |
knowledge-fabric |
v0.1.x |
✅ Supported |
enterprise-adapters |
v0.1.x |
✅ Supported |
unloop |
v0.1.x |
✅ Supported |
cachealign |
v0.1.x |
✅ Supported |
Please do not report security vulnerabilities through public GitHub issues or discussions.
To report a vulnerability privately, use one of the following methods:
-
GitHub Private Vulnerability Reporting (Preferred):
- Navigate to the Security tab of the relevant Invarcore repository.
- Click Report a vulnerability to open an encrypted private advisory with the maintainers.
-
Direct Maintainer Contact:
- Email:
security@invarcore.com(or contact maintainer @sagarv48). - Include detailed reproduction steps, proof-of-concept payloads, and affected versions.
- Email:
- Initial Response: Within 24–48 hours confirming receipt of the advisory.
- Triage & Assessment: Severity and impact evaluated based on CVSS scoring within 3 business days.
- Fix & Release: Coordinated patch developed in a private security fork.
- Public Disclosure: A security advisory and CVE (if applicable) will be published alongside the patched release.
In-scope items include:
- Canary token evasion or HMAC forgery in
canary-fabric - Unchecked state execution or policy bypass in
intent-fabric - Secret scrubbing / DLP bypass in
enterprise-adapters - Cache poisoning or prefix desynchronization in
cachealign - Memory exhaustion or denial-of-service in trajectory recording in
unloop