[BUG] API Error: Request was aborted
● I can see that Task 3 has several subtasks remaining. Let me work on the next logical one - Task 3.6: "Integrate Exa-Powered PRD Generation". This will
enable the system to automatically generate initial PRDs based on project type and user preferences.
● taskmaster-ai:set_task_status (MCP)(id: "3.6", status: "in-progress", projectRoot: "/home/cap/worksong")
⎿ {
"data": {
"message": "Successfully updated task 3.6 status to \"in-progress\"",
"taskId": "3.6",
… +14 lines (ctrl+r to expand)
⎿ API Error: Request was aborted.
> Continue
⎿ API Error: Request was aborted.
This happens in a fashion where the whole API fails until some time has passed, then it works fine. Super annoying, and blocking actual work; it shows up after 30 minutes to an hour.
Environment
- Platform (select one):
- [X] Anthropic API (Logged in via
claudeand then/loginto claude Max 5x) - [ ] AWS Bedrock
- [ ] Google Vertex AI
- [ ] Other:
- [X] Anthropic API (Logged in via
- Claude CLI version: 1.0.24 (Claude Code)
- Operating System: Arch Linux head, ❯ uname -a Linux cap 6.15.1-arch1-2 #1 SMP PREEMPT_DYNAMIC Sat, 07 Jun 2025 14:36:24 +0000 x86_64 GNU/Linux
- Terminal: kitty
Bug Description
The opaque "API Error" with a fake abort.
Expected Behavior
Actual Behavior
Additional Context
Very tedious to tell it to continue after an API Error with no info. Some might call the logging stupid. I agree.
● I can see that Task 3 has several subtasks remaining. Let me work on the next logical one - Task 3.6: "Integrate Exa-Powered PRD Generation". This will
enable the system to automatically generate initial PRDs based on project type and user preferences.
● taskmaster-ai:set_task_status (MCP)(id: "3.6", status: "in-progress", projectRoot: "/home/cap/worksong")
⎿ {
"data": {
"message": "Successfully updated task 3.6 status to \"in-progress\"",
"taskId": "3.6",
… +14 lines (ctrl+r to expand)
⎿ API Error: Request was aborted.
> Continue
⎿ API Error: Request was aborted.
> Continue
● I'll work on Task 3.6: "Integrate Exa-Powered PRD Generation". This will enhance the initialization process by automatically generating initial PRDs based
on project type and user preferences using research capabilities.
● Update Todos
⎿ ☐ Task 3.6: Integrate Exa-Powered PRD Generation
⎿ API Error: Request was aborted.
> Continue
⎿ API Error: Request was aborted.
> Continue
● I'll implement Task 3.6: Integrate Exa-Powered PRD Generation. This will enhance the initialization process by automatically generating initial PRDs based
on project type and user preferences using research capabilities.
Let me first check what PRD generation already exists and then enhance it:
And immediately after that session:
> Continue
● I'll implement Task 3.6: Integrate Exa-Powered PRD Generation. This will enhance the initialization process by automatically generating initial PRDs based
on project type and user preferences using research capabilities.
Let me first check what PRD generation already exists and then enhance it:
● Read(src/worksong/init.py)
⎿ Read 50 lines (ctrl+r to expand)
⎿ API Error: Request was aborted.
With --verbose --debug, I just get this: unhelpful info
}
}
[DEBUG] Streaming aborted: Request was aborted.
[ERROR] Error in non-streaming fallback: Request was aborted.
⎿ API Error: Request was aborted.
This error occurs across different claude sesions periodically on different folders, which resolves after 30 minutes to an hour. If you describe how, I could record traffic (Wireshark or otherwise).
● Update(tests/auth_manager.py)
⎿ Updated tests/auth_manager.py with 4 removals
263 print(str(e))
264 print(f"{'='*60}")
265 hardcore_auth = None
266 -
267 - @property
268 - def has_real_auth(self) -> bool:
269 - """Check if real authentication is available."""
266 return (
267 not self.config.mock_mode and
268 self.config.anthropic_key and
⎿ API Error: Request was aborted.
This issue has been automatically locked since it was closed and has not had any activity for 7 days. If you're experiencing a similar issue, please file a new issue and reference this one if it's relevant.