A target uncertified-DAG architecture with independent ML-DSA-65 operator signatures, Rust-to-RISC-V execution, and a plaintext mempool. Four-second finality, native modules, and economic execution are not inferred from the stale development RPC.
Monolythium's design targets deterministic finality through a DAG that avoids a per-block certification round, independent ML-DSA-65 operator signatures, and Rust-to-RISC-V execution. These are architecture claims, not current performance claims.
The profile combines Starfish-C, ML-DSA-65, MDBX, BLAKE3, and libp2p. The assembly still needs public release evidence. Spending policy remains a target; the in-tree bridge, confidential-value construction, and SLH-DSA emergency registry were removed.
The intended protocol avoids per-block certification round-trips and uses Reed-Solomon erasure with independent ML-DSA-65 operator signatures. O(Mn) is a design analysis, not a measured current-network throughput claim.
The target certificate is a 7-of-10 bitmap over independent ML-DSA-65 (FIPS-204) operator signatures, with no shared signing key, aggregate signature, or key ceremony. The stale development head does not demonstrate finality or a four-second cadence.
The intended artifact path compiles Rust to deterministic, cycle-metered RISC-V. Token, market, policy, and agent-commerce capabilities require individual release and activation evidence. Monolythium is not EVM-compatible at execution, and no in-tree bridge is active.
The protocol profile uses a standard plaintext mempool rather than an encrypted lane. Deterministic inclusion ordering depends on advancing, finalized consensus and is not evidenced by the stale public RPC. Any external-provider route must disclose its own controls and trust model.
The native transaction-envelope profile uses ML-DSA-65 (NIST FIPS-204). The SLH-DSA emergency backup and registry were removed; future cross-family crypto agility requires a new genesis and independent review. This is not a whole-stack post-quantum claim.
The documented user-facing format is bech32m with HRP `mono`, wrapping a compact account identifier. Wallet recovery is standard BIP-39 to ML-DSA-65; PQM-1 was removed.
The proposed registry groups human, agent, cluster, contract, and system names. Registration, pricing, reservations, and activation must be verified from a versioned release; this page does not present them as current network state.
No confidential-value capability is current. A replacement needs a new post-quantum design and independent review.
The earlier Pedersen/Bulletproofs confidential-value construction was removed. A future privacy denomination would require a new post-quantum design, conservation evidence, independent review, and explicit activation. No private LYTH capability is current.
The implementation profile describes MDBX-backed state and a binary Merkle representation over BLAKE3. Public source review, advancing-network evidence, and reproducible tests remain release gates. Stateless proofs are future research, not a current capability.
The network design uses libp2p gossip. The in-tree bridge was removed. Any future cross-chain route must be shipped separately by a provider or issuer with verification, limits, and trust assumptions visible to users.
Attestation, consent, issuer registry, discovery, availability, escrow, spending policy, and reputation remain target capabilities unless a versioned release proves otherwise. Stele is a separately release-gated standalone web product; it does not activate this suite.
The public RPC was stale at height 74,907 when checked on 16 July 2026.
Widgets report the latest response and label stale or unavailable values; they do not certify finality.