github / github/copilot-cli

Background subagent can remain running indefinitely after tool activity stops

Open
#4,850 2 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

triage
Dominant language
Shell
Stars
11.2k
Forks
1.9k
Avg merge
14h 16m
Merged PRs (30d)
6

Description

Describe the bug

A background review subagent remained marked as running long after observable activity stopped. The status stayed at 100 completed tool calls and zero completed turns while elapsed time continued increasing past 15 minutes. The parent session received no result and could not complete the workflow.

Affected version

1.0.83

Steps to reproduce the behavior
  1. Start a background general-purpose subagent for a final branch review.
  2. Allow it to perform tool calls.
  3. Observe the agent reach approximately 100 completed tool calls.
  4. Continue checking its status after several minutes.
  5. The agent remains running, with the tool-call count unchanged and no completed turn or result.
Expected behavior

The runtime should detect stalled subagents, surface the blocked tool or execution state, and either recover, fail with an actionable error, or enforce a bounded timeout. The parent should not wait indefinitely with no progress signal.

Additional context
  • Linux
  • Copilot CLI 1.0.83
  • Observed elapsed time: over 15 minutes
  • Status remained: running, 100 tool calls completed, 0 turns completed
  • Implementation and the full 78-test regression suite had already completed; the stalled operation was only a final read-only review.
  • No private repository content or session logs are attached.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Reproduce the background final-branch review on Linux and monitor it after tool activity reaches about 100 calls. Start by tracing the subagent status and parent-wait behavior; the full 78-test regression suite is available for validation. Done means stalled work is detected, an actionable outcome is surfaced, and the parent no longer waits indefinitely.

Written by the indexing model from the issue text.

Assessment

Tech stack
shell
Domain
cli
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
52/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.