openai / openai/openai-agents-python
Local shell executor failure aborts the whole run instead of reaching the model
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 29.6k
- Forks
- 4.8k
- Avg merge
- 1d 20h
- Merged PRs (30d)
- 123
Description
Hey! Found that when the local shell executor itself fails (missing binary, sandbox denial, failing command with check), the raw error propagates out of Runner.run and kills the whole run — the model never sees the failure and can't retry. The sibling executors degrade to model-visible failed output instead. Happy to PR a fix that surfaces the failure as output like the others if that helps.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start at Runner.run and trace the local shell executor failure path, then compare it with the sibling executors that already expose failures to the model. Done means missing binaries, sandbox denials, and checked command failures become model-visible failed output so the run can continue instead of aborting.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, shell
- Domain
- ai, tooling
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 72/100