If you believe you have found a security vulnerability in @saihm/client-pro (the SAIHM client-side cryptography library), please report it privately so that we can investigate and remediate before public disclosure.

Private channel: architect@saihm.coti.global

Please include, where possible:

• A description of the vulnerability and its potential impact
• Steps to reproduce, or a proof-of-concept
• Affected version(s) of @saihm/client-pro
• Whether the issue is in the sealing/identity/sharing crypto, the canonical wire serialization, or in a dependency
• Your name or handle if you wish to be credited in the fix

We acknowledge reports within 14 days. We aim to provide an initial assessment and a fix or mitigation plan within 30 days for confirmed vulnerabilities, depending on severity and complexity.

In scope:

• The published npm package @saihm/client-pro and its source in this repository
• The cryptographic envelope: identity derivation, per-cell AES-256-GCM sealing, ML-KEM-768 authenticated sharing, signature verification, and the canonical wire serialization

Out of scope (please report to the relevant project instead):

• Vulnerabilities in third-party MCP clients (Claude Code, Claude Desktop, Cursor, etc.) — report to the client vendor
• Vulnerabilities in the underlying Model Context Protocol — report to https://github.com/modelcontextprotocol
• Vulnerabilities in the COTI V2 blockchain network — report to COTI Group
• Vulnerabilities in unrelated open-source dependencies — please report upstream and let us know so we can pull a patched version

We follow a coordinated-disclosure model. Once a fix or mitigation is available we will:

1. Release a patched version of @saihm/client-pro to npm
2. Publish a security advisory on the GitHub repository
3. Credit the reporter (with permission) in the advisory and release notes

@saihm/client-pro performs every operation that touches plaintext or the master secret on the client. The primitives are:

• ML-DSA-65 (NIST FIPS-204) for post-quantum agent identity and signing
• ML-KEM-768 (NIST FIPS-203) for post-quantum authenticated sharing
• AES-256-GCM for per-cell confidentiality
• HKDF (RFC 5869) for key derivation
• Implementations from the audited @noble/* family

If a vulnerability is in the protocol specification itself rather than this implementation, please indicate that in your report.

Responsible disclosure protects the broader agent ecosystem. We appreciate the time and care of security researchers who report issues to us privately.