Research / What good looks like

What “good” looks like

Leaving ECC is not one switch. The postures on this site are a compressed version of a real engineering ladder.

Post-quantum native

The chain never used ECC for spend authority. Quantus, for example, signs with NIST ML-DSA (FIPS 204) and uses ML-KEM on the wire. QRL used XMSS from genesis. There is no “migrate the account model” problem because the account model was never elliptic-curve.

Hybrid / transitioning

Production still has ECC in the critical path, but PQ is not a blog post. Algorand is the current large-L1 example: Falcon-1024 State Proofs since 2022, Falcon-authorized mainnet transactions in 2025, and a public roadmap for native PQ accounts with consensus still on Ed25519 / ECC-VRF. That is a 40–70 weight, not a 0. Nervos CKB’s opt-in SPHINCS+ locks sit in the same band: available, not default.

Key rotation and harvest-now-decrypt-later

Even a perfect PQ signature scheme fails if old ECC public keys remain spendable forever, or if ciphertext encrypted to ECC long-term keys was stored by an adversary. Rotation, migration windows, and (where relevant) re-encryption are part of “good.” Project Eleven’s post-Q-Day ownership proofs are a last-resort tool for coins that miss the window, not a substitute for moving.

What does not move the score

• A research repo with no mainnet path
• A roadmap with no date and no deployed primitive
• ZK proofs that still verify against ECC accounts
• Marketing that says “quantum-resistant” without a scheme and a spec

For how those rules hit fund scores, see the methodology.