HSM / PKCS#11 / Hybrid Signature Spectrums

What you will do: Choose Concatenation, Nesting or Shared-binding model (Silithium-inspired), then Generate Key Pairs, Sign and Verify. Read the full-binding and raw-message EC/ML checks. The local model is not the published Silithium scheme, and this page does not claim strong non-separability or a security proof.

Worked example: With a valid local-model sample, both full-binding components pass while the separate raw-message checks do not. Recombination Attack retains the EC public key and original EC bytes while signing a replacement ML component with its own key: inspect the actual EC, ML and combined results.

Runtime and privacy: The cryptographic exercise runs in this browser. Review the site privacy terms before entering sensitive material; use synthetic inputs for learning and evaluation.

Browse all Crypto Lab tools · Learn with Hybrid Cryptography

For your role

Developer / Engineer
Choose Concatenation, Nesting or Shared-binding model, then Generate Key Pairs, Sign and Verify. Compare the actual full-binding and raw-message EC/ML checks. Concatenation and nesting use the PKCS#11 trace for ML-DSA; the local shared-binding model runs in JavaScript and is not published Silithium or a security proof.
Security Architect
Compare the observed component checks and the public-only ML replacement experiment. The local shared-binding model is not published Silithium or a non-separability proof. For a production certificate decision, require a versioned standard/profile, relying-party compatibility checks and implementation evidence from your own tools.
Researcher / Academic
Sign the same message under all three options and compare measured layouts and actual checks. The local shared-binding model uses fresh random EC nonces; its public-only ML replacement experiment tests one binding case, not the published Silithium construction or a general non-separability property.
Certification & Validation Engineer
Sign one message under Concatenation, Nesting and Shared-binding model. The first two expose ML-DSA calls in the PKCS#11 trace; the local model uses provider-internal JavaScript operations outside that boundary. Its component/replacement outcomes are not published-scheme conformance or validation evidence. Evaluate the exact versioned standard/profile and implementation evidence with your production tools.

Hybrid Signature Spectrums

Compare concatenation, nesting, and a local shared-binding model. Observe component checks and a public-only replacement experiment; these do not establish a security proof.

softhsmv3 WASM ready · CKM_ML_DSA (PKCS#11 v3.2) active for concatenation / nesting
Why hybrid signatures? — Migration context

During the PQC transition, systems must support verifiers that understand only classical algorithms (legacy) and verifiers that require post-quantum algorithms (future-safe). Hybrid signatures let one signed artifact satisfy both.

The critical security question is: what prevents an attacker from stripping the PQC component? If they can, the document is no longer quantum-safe — it degrades silently to classical security. This tool explores that question with component checks and a limited local binding experiment.

Standards: IETF RFC 10073 (a labelled concatenation, not the bare form shown here); RFC 9955, Hybrid Signature Spectrums (the concatenation, nesting and fused categories); ePrint 2025/2059 (Silithium). Profile distinction: RFC 10073 specifies Composite ML-DSA for X.509 PKI, whereas RFC 10082 specifies Composite ML-DSA for CMS / S/MIME.

Concatenationsig₁ ‖ sig₂Non-separability: None / WNS

Bare concatenation, shown here with no label or prefix (RFC 9955 Table 2, case 1): either component can be stripped and verified alone. RFC 9955 (section 1.3.4) notes that concatenation with a label in the message gives weak non-separability; Composite ML-DSA (RFC 10073 section 9.2.3) works that way.

RFC 9955 §2
ML-DSA-65 — softhsmv3 WASM (CKM_ML_DSA, PKCS#11 v3.2)EC-Schnorr — @noble/curves (no PKCS#11 Schnorr)
Educational sandbox demonstration. Keys are ephemeral in browser memory (SoftHSM3 WASM and @noble) — this is not a FIPS 140-3 / CMVP-validated production HSM or an interoperable RFC 10073 / RFC 10082 implementation. The local shared-binding model uses fresh random EC signing nonces and is inspired by ePrint 2025/2059 (Devevey, Guerreau, Roméas). Its approximation uses ML-DSA external-μ operations; it is not the published scheme or its security proof. Validate target profiles, HSM support, and relying-party verification in your production environment.
Production Tool Capabilities & Verification Requirements

What this sample check establishes: Demonstrates composite and dual classical + PQC digital signature generation and verification where both `ECDSA`/`Ed25519` and `ML-DSA` must verify.

Browser sandbox limitation (non-certification boundary): Executes in browser memory; it does not test third-party X.509 path-validation engines or legacy middlebox ASN.1 parsers.

What to require and test in your production tools:

  • IETF RFC 10073 (Composite ML-DSA) OID support and strict AND-verification semantics (both component signatures must pass).
  • Prefix/Label domain separation as defined in RFC 10073 §2.2 and §9.4, and no reuse of component keys outside the composite (RFC 10073 §9.4 MUST).
  • Negative test verification confirming that corrupting either the classical or the ML-DSA component causes overall verification failure.

Try it

  1. 1. Predict before you run it.
  2. 2. Run or change the tool above.
  3. 3. Explain what the result shows.

In this sample, which construction verifies BOTH raw-message component signatures independently?

A verifier expects a hybrid signature under a both-components policy. It verifies the classical signature but the post-quantum signature fails. What should the application report?

Next step

Turn it into a plan: Hybrid Transition Planner

This tool practises the Hybrid Cryptography module, phase 5 (Pilots & Migration); Hybrid Transition Planner produces a deliverable of that phase.