anyswe_claude_code: responses_api_agents: anyswe_agent: entrypoint: app.py domain: coding description: SWE-bench run by Claude Code natively inside the task container. value: Eval software engineering capabilities on SWE-bench with any Gym agent. agent_server_module: responses_api_agents.claude_code_agent.app agent_server_class: ClaudeCodeAgent agent_config_class: ClaudeCodeAgentConfig agent_kwargs: model: model max_turns: 100 timeout: 1800 system_prompt: "You are fixing a bug in the repository at /testbed. Edit the source files to fix the root cause, then verify. You MUST actually apply the fix with your file-editing tools; do not just describe it." container_formatter: ${oc.env:ANYSWE_CONTAINER_FORMATTER} sandbox_provider: sandbox sandbox_model_base_url: ${oc.env:NEMO_GYM_SANDBOX_MODEL_BASE_URL,null} swebench_agent_timeout: 2700 swebench_tests_timeout: 1800 concurrency: 4 model_server: name: policy_model type: responses_api_models datasets: - name: example type: example jsonl_fpath: responses_api_agents/anyswe_agent/data/example.jsonl anyswe_claude_code_environment_server: environment_servers: legacy_agent: entrypoint: app.py agent_server: type: responses_api_agents name: anyswe_claude_code