Guided website changes and verification
Three separate pages were added: Getting Started, the three-step real proof experience and launch preparation. Home links now explain the next action before opening its tools.
Launch preparation workspace
Five steps organize project information, issuance parameters, allocation and vesting, liquidity and reviews, and next priorities. Formal supply and allocation percentages are not inferred from test defaults. Undecided fields may stay empty. An example purpose statement is explicitly labeled for editing.
A draft is written to this browser only after Save or Continue passes validation. It is not uploaded. Refresh recovery, text export, JSON backup and draft clearing are supported. This is not a cloud account and does not issue tokens, pay, create pools or apply for listings.
Real proof experience
The three-step flow reuses actual Semaphore/Groth16 verification: create synthetic identities and a proof, check session replay protection and application separation, then explain and export the results. The record is not an identity document or certificate. Refreshing or clearing resets the session; network-wide replay protection is not provided.
Verification results
Eighteen recorded code tests passed, including parent tests: twelve existing real ZK and TIME/local Uniswap V2 tests plus six launch draft engine tests. Added checks covered integer supply, public EVM addresses, rejection of private-key-shaped input, exact allocation percentages, over-allocation, invalid URLs/email, undecided fields and honest export status.
Browser checks completed the five-step draft wizard, invalid-format rejection, unresolved-field display, text-only rendering of user input, both actual file downloads, refresh recovery and application-only draft clearing. The experience generated and verified a real proof, checked replay and application separation, and exported a record.
At a simulated 390px mobile viewport, all three pages had usable menus and layouts without horizontal overflow. Empty fields could continue but remained unresolved; the example did not save automatically. Real proof computation also ran at that layout size. This was not a physical-phone performance test.
Browser checks also verified integer precision: a 16-digit integer was retained exactly, while values exceeding the 18-decimal ERC-20 range were rejected rather than silently truncated.
Screenshot capture timed out; generated images were not presented as real screenshots. Page interactions and semantic checks completed. Independent audit, multi-device sync, real identity registration, public-testnet wallet acceptance, mainnet issuance and exchange applications remain incomplete.