Status: Self-reviewed by the Ciphera team; not yet independently audited. This is open-source cryptographic software — review the code and the threat model (THREAT-MODEL.md) before relying on it.

Report security issues privately to security@ciphera.net. Please do not open a public issue for a suspected vulnerability.

Include, where possible: affected component (ciphera-tessera core/sidecar, tessera-go, or @ciphera-net/tessera) and version/commit, a description and impact assessment, and a minimal reproduction. PGP-encrypted reports are welcome; the key fingerprint is published alongside the security@ address.

• Acknowledgement within 3 business days.
• Triage + initial assessment within 10 business days, with a severity and a remediation plan.
• Coordinated disclosure: we request a 90-day embargo from first report to public disclosure, or until a fix ships — whichever is sooner — and will coordinate timing with you.
• Credit: we credit reporters in the advisory unless you prefer to remain anonymous.

We will not pursue or support legal action against good-faith security research that: respects the embargo; avoids privacy violations, data destruction, and service degradation; only interacts with accounts/data you own or have explicit permission to test; and gives us a reasonable window to remediate before disclosure. Good-faith research conducted under this policy is authorised.

Until the first stable release, only the latest main of each repository is supported. After a stable release, the most recent minor series receives security fixes.

In scope: the cryptographic core and protocol (ciphera-tessera), the Go server SDK (tessera-go), the browser SDK (@ciphera-net/tessera), and the pinned wire/crypto contract (see SPEC.md).

Out of scope / known limits: the residual limits documented in THREAT-MODEL.md are known and accepted for v1 (e.g. non-extractable keys are an API-layer guard, not process isolation; constant-time execution is not fully achievable in TS/WASM; the recovery phrase is an immutable, un-zeroable string). Reporting one of these documented limits is welcome but will be triaged as a known limitation rather than a new vulnerability.