loopback · v1.0.0
DeepSeek V4 Flash
The model produced a recognizable offline puzzle game and reported all checks passing, but independent fresh-browser validation could not complete Level 1 at its stated par. Solvability, win transitions, sequential unlocking, persistence, and the claimed 12-level completion chain are therefore not established.
Run the artifactStarts only after your click, inside an opaque-origin sandbox with network and site access blocked.
Sandbox active. Click inside the game before using keyboard controls.
Artifact hashes
ba0f765cee16aa03d56f84ba401a2f21a5621011b112e5f1dda69df42e22c623Criterion outcomes
boot-offline
Loaded from file:// in a fresh Chromium context with only the document request and no external URL, fetch, XHR, WebSocket, or import references.
core-loop
Pointer and keyboard rotation, undo, reset, and locked-level behavior worked. Exhaustive testing of all 81 ordered two-click sequences for Level 1 produced no completion at the stated par, so the core progression loop did not pass.
levels-solvable
The model-authored verification log claims 12/12 levels, but an independent fresh-browser run could not establish even the Level 1 completion needed to reach the remaining levels.
persistence-recovery
localStorage code is present, but completion and unlock state could not be created through the tested UI, so reload recovery was not established.
keyboard-complete
Arrow navigation, Enter rotation, focus indication, named controls, and live-region structure worked. Keyboard-only completion of Levels 1 through 3 could not be demonstrated because Level 1 did not reach a win state.
Dispute finding
The central success claim did not survive independent validation. The preserved record shows which checks still passed and where progression failed.