Ecosystems · Sui

Sui: the foundations, stated without inflation.

The graph intelligence and adapter architecture that a Sui implementation needs exist and are tested. The decoding that a Sui pre-sign decision would require does not.

What exists

Graph work first, chain work second.

Argus began as graph intelligence, and that part is chain-agnostic: canonical subjects, provenance and control relations, cycle-safe traversal, blast radius and coverage-aware inheritance. None of it assumes an EVM chain, which is exactly why a second ecosystem is tractable.

The adapter boundary that a Sui implementation would sit behind is defined, tested and enforced. What is missing is the Sui-specific half.

What is missing

Programmable-transaction-block decoding, without which a Sui intent cannot be read exactly.
A Sui structural discovery implementation for provenance and control relations.
Any connected intelligence source for Sui subjects.

TodayA partial decode is worse than no decode. Until PTBs can be read exactly, Argus makes no Sui pre-sign claim at all.

Sui capability status

Architecture ready is not the same as shipped.

Structural graph modelProven offline
Canonical subjects, relations, cycle-safe traversal and blast-radius calculation — the graph work the Sui track is built on, independent of any one chain.
Adapter architectureArchitecture ready
The adapter contract, capability declaration and injected-transport discipline are defined and enforced by architecture tests.
Sui subject identityArchitecture ready
Canonical subject identifiers are chain-agnostic by construction, so Sui objects and addresses have a representation to attach evidence to.
PTB pre-sign decodingNot available
Programmable transaction blocks are not decoded. Without exact decoding there is no honest pre-sign decision for a Sui transaction, and Argus will not approximate one.
Sui screening pathNot available
No end-to-end Sui pre-sign screening exists. Nothing on this site should be read as offering one.
Live Sui indexingPlanned
Sequenced behind PTB decoding: indexing structure is only useful once the transaction being assessed can be read exactly.

Proven offlineImplemented and covered by automated tests against fixtures or injected transports. No live endpoint involved.

Architecture readyThe contracts and seams exist and are tested; the specific implementation behind them has not been written.

PlannedScoped and sequenced, not yet built.

Not availableDoes not exist. No partial implementation to point at.

Why it is sequenced this way

Depth on one ecosystem before breadth across several.

Proving the full path once — decode, structure, policy, verdict, disposition, receipt — is worth more than shallow support across several chains. The Arc track carries that proof today.

The architecture does not privilege it. When Sui decoding is implemented, it plugs into the same engine, produces the same shape of decision, and carries the same reasoning structure.

Where Sui sits

The engine treats Sui as one more ecosystem behind the adapter contract, not a separate product.
Graph and coverage semantics are already shared, so no parallel decision logic is needed.
Resuming Sui work is an implementation task, not an architectural change.

Interested in the Sui track?

If Sui pre-sign decisioning matters to your platform, the architecture and the specific gap are worth a direct conversation.

Start a conversation Arc track