SWE-agent
SWE-agent copied to clipboard
princeton-nlp
SWE-agent takes a GitHub issue and tries to automatically fix it, using GPT-4, or your LM of choice. It solves 12.47% of bugs in the SWE-bench evaluation set and takes just 1 minute to run.
This project is pretty great BUT we need more options to use different LLM's.You don't have to worry about creating a solution which supports 100+ LLM easily as LiteLLM is...
Example PR: https://github.com/klieret/swe-agent-test-repo/pull/12 ## Safeguards * PR is only created as draft and has to be manually marked as "ready for review" PRs only get created if * swe-agent reports...
It would be great if it could run on a local directory and suggest a new git branch so I could compare the changes + request a change without needing...
re: https://github.com/princeton-nlp/SWE-agent/issues/60
Looking at the current 'search' command that the agent has access to, it seems it just uses `grep` for searches within a file, and `find` for searches by filename: https://github.com/princeton-nlp/SWE-agent/blob/79038832dfda12e29e1a4ab19d3ee3c8f86e2b9d/config/commands/search.sh#L89-L95...
Happy to contribute here, I'm sure others are already thinking here but figured I'd write some notes about what would've made trying this out easier: - A locally served onboarding...
It would be great to have the source code language be configurable beyond just python. It would involve tweaking some of the prompts https://github.com/princeton-nlp/SWE-agent/blob/3d8a6c44f16f190735c774676fea80315cd84da5/config/default.yaml#L38 and also changing the shell scripts...
Automatically process SWE-agent repo issues. Investing effort in this more complex feature will accelerate/leverage future development of SWE-agent itself.
Metadata
Owner
Metadata
SWE-agent takes a GitHub issue and tries to automatically fix it, using GPT-4, or your LM of choice. It solves 12.47% of bugs in the SWE-bench evaluation set and takes just 1 minute to run.