Changelog

2025-10-02

• Added Bash tool and grep tool. #209 #218
• Added debugging trajectory viewer. #233 #236
• Removed dependency on local filesystem allowing to run without mounting volume in Docker. #228
• Added tutorial notebook. #213
• Added R2E-Gym as a new benchmark. #232
• Added KubernetesTerminal. #234

2025-06-13

Added scripts to filter trajectories generated by debug-gym.

2025-06-11

Added support to SWE-smith. Users can use the tasks shipped with the official SWE-smith package, or customized tasks generated using SWE-smith.

2025-06-03

• Refactored debug_gym/agents/llm_api.py into separate modules in debug_gym/llms/ for OpenAI, AzureOpenAI, Anthropic APIs, and human mode, allowing for easier extension to other LLM providers in the future.
• Improved the Human mode to support better prompt completion and error handling.

2025-05-28

Improved the View tool, added the start and end arguments so the agent can specify a particular chunk of code to view.

2025-05-22

Added in the analysis folder a Flask app to view .jsonl log files in the browser.

2025-05-20

Changed the tool-calling syntax to be compatible with the OpenAI and Anthropic function-calling formats.

• Switched tools (view, edit, pdb, listdir, eval) to a function-call API with explicit arguments and environment injection.
• Overhauled LLM interfaces to define, parse, and format function calls, and updated agents to consume ToolCall objects.
• Removed the old conversational-prompt flag from configs.