{"schemaVersion":"agentdepot.eval_plan.v1","asOf":"2026-10-28","storesIpAddresses":false,"query":"tool-use","matchMode":"matched_task","taskType":"tool-use","label":"Function-Calling, MCP Tool-Use & Multi-Step API Agents","recommendedBenchmarks":[{"id":"bfcl","name":"Berkeley Function-Calling Leaderboard (BFCL)","officialUrl":"https://gorilla.cs.berkeley.edu/leaderboard.html","graderType":"AST structural match + live execution check"},{"id":"tau-bench","name":"τ-bench (Tau-bench)","officialUrl":"https://github.com/sierra-research/tau-bench","graderType":"Final DB state + policy pass^k check"}],"depotBenchmarkSuite":{"type":"tool_use","tasksApi":"/api/benchmark/tasks?type=tool_use","submitApi":"/api/benchmark/runs"},"stackPicks":{"evalHarness":{"jtbdId":"agent-trajectory-tool-use-evals","pickProviderId":"braintrust","coPickProviderId":"langsmith","runnersUpIds":["langsmith","arize-phoenix"]},"sandboxOrRunner":{"jtbdId":"public-agent-benchmarks","pickProviderId":"promptfoo","coPickProviderId":"braintrust","runnersUpIds":["braintrust","e2b"]},"ciGate":{"jtbdId":"continuous-eval-ci","pickProviderId":"braintrust","coPickProviderId":"promptfoo","runnersUpIds":["promptfoo","confident-ai"]},"judgeCalibration":{"jtbdId":"llm-as-judge-calibration","pickProviderId":"braintrust","coPickProviderId":"patronus","runnersUpIds":["patronus","ragas"]}},"targetMetrics":["AST schema validity %","irrelevance abstention accuracy","tool-sequence optimality","costPerSolvedTaskUsd"],"antiContaminationRule":"Include irrelevance turns (where no tool should be called) and missing-parameter turns so eager tool callers fail honestly.","nextCalls":["/api/decide?jtbd=agent-trajectory-tool-use-evals","/api/benchmark/tasks?type=tool_use","/api/playbooks/agent-evals"]}