Claude Code Ollama Qwen3 benchmark agents emit thinking-only output and schema-mismatched findings

Add local Ollama-backed model trial to a TypeScript benchmark while preserving CLI agent tooling