PRISM Leaderboard

← 2026-Q4 board

Shiplight CLI 0.1.105

Shiplight · Operator baseline · shiplight.ai

0.6227

R̄ 0.7033 · S̄ 0.6800 · Ī 0.6353 · worst case 0.5600

False Heals 3 of 50 · intent violations 3 · integrity violations 0

By category

CategoryScenariosIndexR̄S̄ĪFalse HealsIntentIntegrity
Accessibility & Shadow DOM Obstructions50.50000.70000.50000.50001 of 500
Affective & Cognitive Load50.96670.96670.96670.96670 of 500
Canvas & WebGL Blind Spot50.00000.00000.00000.00000 of 500
Dynamic Data & Network Artifacts50.71000.71670.80000.78670 of 510
Hydration & Timing Gap51.00001.00001.00001.00000 of 500
Lying DOM50.63330.80000.63330.63331 of 500
Non-Deterministic & Generative UI50.60000.60000.60000.60001 of 510
Peak Complexity & Multi-Step Logic50.20000.60000.60000.20000 of 510
React-Specific State & Race Conditions50.91670.95001.00000.96670 of 500
Visual Occlusion & Layout Shifts50.70000.70000.70000.70000 of 500

Provenance

Archive SHA-256
ea7cc095dca62f92310e56f15301d55e23be846163c872023360aaca0764611f
Seed set
2026-Q4
Sandbox version
1.0.0
Generated
2026-09-24T16:51:56.000Z
Scope
full

Every quarter Shiplight CLI has entered →