Skip to content

Error messages from model provider during a subagent don't make way to primary agent #5204

Closed as not planned
kiennq/opencode
#4

Description

@bcardarella

Description

I hit my weekly Claude limit and the primary agent just continued to spin over and over until I finally checked the subagent and saw the rate limit exceeded error from Anthropic.

OpenCode version

No response

Steps to reproduce

No response

Screenshot and/or share link

No response

Operating System

No response

Terminal

No response

Activity

  1. github-actions commented on Dec 7, 2025

    @github-actions
    Contributor

    This issue might be a duplicate of existing issues. Please check:

    Feel free to ignore if none of these address your specific case.

  2. Haohao-end commented on Mar 21, 2026

    @Haohao-end

    I’ve opened a fix for this.

    Summary:

    • TaskTool.execute() was treating a failed child assistant as a successful task result
    • child info.error was ignored, so provider failures could be downgraded into an empty <task_result>
    • the fix now throws on child info.error, which lets the existing parent/orchestrator error path surface the real provider message

    Validation:

    • added a success-path regression test
    • added a provider-error propagation regression test
    • verified with bun test test/tool/task.test.ts
  3. removed
    bugSomething isn't working
    on May 3, 2026
  4. github-actions commented on Jul 3, 2026

    @github-actions
    Contributor

    To stay organized issues are automatically closed after 60 days of no activity. If the issue is still relevant please open a new one.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions