Please do not open public GitHub issues for security reports.
Email: security@identities.ai
We commit to:
- Acknowledging receipt within 48 hours.
- Providing an initial severity assessment within 5 business days.
- Keeping you updated on remediation progress.
- Crediting you in the advisory unless you prefer otherwise.
| Version | Supported |
|---|---|
v1.0.0-alpha.* |
|
< v1.0.0-alpha.1 |
❌ Not supported |
Once v1.0.0 stable is released, the support policy will specify which major/minor versions receive backported security fixes.
- We follow a 90-day embargo from the initial report to public disclosure as the default.
- For actively-exploited vulnerabilities, we may shorten the embargo after coordinated patching.
- The full coordinated advisory (CVE, scope, impact, mitigation) will be published in the
security/directory of this repository and on our public advisories channel.
In scope:
- Cryptographic correctness of the reference implementation (Go) and published SDKs (TypeScript, others as released).
- The canonical serialization implementation — any interop-breaking behavior where implementations produce divergent signable bytes.
- The verifier algorithm — any case where
Verify()returns an incorrect result. - The scope-vocabulary implementation — any privilege-escalation path.
- The
cmd/ratifyCLI's handling of private key material. - Resource-exhaustion or DoS against the verifier.
Out of scope (report to the relevant party, not us):
- Vulnerabilities in third-party dependencies that are already known and patched upstream. Report to the dependency, then open a PR here to bump the version.
- Operational issues at any Identities AI-hosted service (Registry, Verify API). Those are reported via
security@identities.aiwith the service name in the subject line — a separate handling path. - Attacks against out-of-protocol concerns listed in
docs/EXPLAINED.md§5.2 (endpoint malware, social engineering of delegators, etc.). These are real risks but are not protocol defects.
| Scenario | Handling |
|---|---|
| You find a way to verify a maliciously-crafted proof bundle | In scope — email us first. Classic responsible disclosure. |
| You find a way to produce canonical JSON bytes that differ from the Go reference for a valid v1 data model | In scope — email us first. Interop breaks trust. |
| You notice the recommended challenge window is too long | Not a vulnerability — open a GitHub issue or PR. Policy discussion is public. |
| You find a logic bug in an Identities AI-hosted service | Email us with [hosted-service] in subject. Separate from the protocol project. |
Reporters who responsibly disclose verified vulnerabilities will be credited here (with consent) once the first advisory is published.