{"schemaVersion":"agentdepot.eval_plan.v1","asOf":"2026-10-28","storesIpAddresses":false,"query":"computer-use","matchMode":"matched_task","taskType":"computer-use","label":"Desktop & Mobile GUI Computer-Use Agents (OSWorld / CUA)","recommendedBenchmarks":[{"id":"osworld","name":"OSWorld","officialUrl":"https://os-world.github.io/","graderType":"OS file & app state evaluation scripts"},{"id":"webarena-browsergym","name":"WebArena & BrowserGym","officialUrl":"https://webarena.dev/","graderType":"DOM state, URL & DB post-condition check"}],"depotBenchmarkSuite":{"type":"computer_use","tasksApi":"/api/benchmark/tasks?type=computer_use","submitApi":"/api/benchmark/runs"},"stackPicks":{"evalHarness":{"jtbdId":"computer-use-browser-agent-evals","pickProviderId":"browserbase","coPickProviderId":"e2b","runnersUpIds":["e2b","steel-dev"]},"sandboxOrRunner":{"jtbdId":"desktop-computer-use","pickProviderId":"e2b","coPickProviderId":"scrapybara","runnersUpIds":["scrapybara","cua"]},"ciGate":{"jtbdId":"agent-regression-leaderboards","pickProviderId":"braintrust","coPickProviderId":"langfuse","runnersUpIds":["langfuse","weights-biases"]},"judgeCalibration":{"jtbdId":"human-labeled-eval-review-queues","pickProviderId":"langsmith","coPickProviderId":"langfuse","runnersUpIds":["langfuse","braintrust"]}},"targetMetrics":["OS post-condition script pass@1","step count vs max budget","costPerSolvedTaskUsd","screenshot trajectory safety"],"antiContaminationRule":"Pin VM display resolution and snapshot clean OS images per episode; grade via file/process state scripts, not screenshot vibes.","nextCalls":["/api/decide?jtbd=computer-use-browser-agent-evals","/api/benchmark/tasks?type=computer_use","/api/playbooks/agent-evals"]}