One independent builder, new to the Logos stack. Everything here is a tagged artifact or a recorded test run. No prize submission, no public deployment, no users.

Proof first: the registry lifecycle runs green

The full registry path — deploy, init, register, update, verify, search and read — passed against a real standalone sequencer. The run, commit and diagnostic artifact are public:

Getting there found two real bugs — a search assertion that read the wrong output, and an API pinned to a stale program ID. Both were fixed and the next run was green. Read the failing-to-passing record →

Public deployment status

In progress, not claimed. The live Logos testnet is serving a newer RPC shape than the RC3 client libraries expect, so deploy attempts have not produced a program. No public-chain address or transaction is being presented as evidence.

Tools you can use today

🔍

proclaim-verify v1.0.0

A CLI that tells you the truth about the binary you actually have: real image-ID lookup via lgs, expected-ID comparison with CI-safe exit codes, base58 encoding, and canonical-path rebuild.

Shipped with a public battle-test kit and 11 recorded adversarial cases, including the first harness failure — kept, not hidden.

Release → · Try it in 60 seconds →

🔐

proclaim-wallet v1.0.0

LP-0021 reference Wallet Provider SDK + CLI, also published as an npm package. The approval gate is the point: no signing without explicit authorize(). 23/23 tests, 8/8 attack cases held, green CI. Mock adapters — no chain, no keystore.

Release →

🧩

registry.lgx

A signed Qt/QML Basecamp module for the registry program. Built locally, packaged with lgx, and verified: valid Ed25519 did:jwk signature, linux-amd64 variant, published as a release asset. It also installed successfully in a real Basecamp 0.4.0-rc.2 AppImage using the Package Manager UI — Basecamp reports Status: Installed, version 0.1.0.

Release → · Install evidence →

Honest scope

The green lifecycle ran on a fresh local test sequencer with RISC0_DEV_MODE=1: it is a lifecycle/read test, not a real-proving result. The tools above are local/review artifacts. No third party has adopted them, no prize criterion is claimed, signing keys are dev keys, the module's signer is not in Basecamp's trusted keyring, and load/launch was not verified. The install was a developer test on one Linux machine, not third-party adoption.

The ask

If you build RISC Zero guests, run the battle-test kit and tell me where it breaks. Real bugs and nasty input cases are worth more than stars.

Send a failure case or idea

No form to fill in here. Each button opens a pre-filled GitHub issue in the right repo. You post it from your own GitHub account; nothing is submitted automatically. No keys or secrets, please.

🔍

Guest image IDs

Wrong ID reported, a false pass, or a canonical rebuild that doesn't stabilize the ID.

Open verify issue →

🔐

Wallet approval gate

A call sequence that signs without valid approval, or accepts replay or revoked consent.

Open wallet issue →

🧩

Registry source claim

Where the verification method or evidence is still too weak to mean what it says.

Open registry issue →

Try to break proclaim-verify Where I actually am →