Tool-calling Traces — OpenAI Format (JSONL)
OpenAI-style tool-call JSONL with function invocation and tool result messages.
LLM tool-calling fixtures
JSONL traces of function/tool calls with arguments and results — for testing agent harnesses and tool routers.
14 of 14 files
Tool-calling Traces — Anthropic Format (JSONL)
Anthropic-style tool_use blocks with paired tool results for agent harness tests.
Tool Call Results (JSONL)
Function-call result rows decoupled from chat messages for router testing.
Multi-step Agent Trace (JSONL)
Multi-step agent action/observation trace for ReAct-style eval harnesses.
OpenAI Tools Schema (JSON)
OpenAI function tools definition JSON paired with the tool-calling traces.
Tool Function JSON Schema
JSON Schema for a sample weather tool's parameters.
Failed Tool Calls (JSONL)
Failed tool invocation rows for agent error-handling and retry tests.
Parallel Tool Calls (JSONL)
Parallel multi-tool call payloads for concurrent tool-router tests.
Tool-calling Mini Trace (JSONL)
Minimal single-turn tool-calling trace for quick parser smoke tests.
Tool Calling — Tool Call Weather
SAMPLE tool-calling JSON (tool-call-weather) for agent harness schema tests.
Tool Calling — Tool Result Weather
SAMPLE tool-calling JSON (tool-result-weather) for agent harness schema tests.
Tool Calling — Tool Call Search
SAMPLE tool-calling JSON (tool-call-search) for agent harness schema tests.
Tool Calling — Tool Parallel Two
SAMPLE tool-calling JSON (tool-parallel-two) for agent harness schema tests.
Tool Calling — Tool Error Unknown
SAMPLE tool-calling JSON (tool-error-unknown) for agent harness schema tests.