hermes - 💡(How to fix) Fix Support real-time bidirectional communication between subagent and main agent

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
RAW_BUFFERClick to expand / collapse

markdown Problem

Currently, when using delegate_task to spawn subagents, the main agent must wait for the subagent to fully complete before receiving any results or status updates. This creates a black-box situation where:

- If a subagent goes down the wrong path, the main agent cannot intervene until it is too late
- No real-time progress visibility — success or failure is only known at the end
- Retries happen blindly — the main agent cannot provide corrective guidance between attempts

Desired Behavior

Enable bidirectional real-time communication between subagent and main agent during execution:

1. Subagent → Main agent: Subagent can push status updates, checkpoints, or partial results mid-flight (e.g., via API calls back to the main agent or a shared message queue)
2. Main agent → Subagent: Main agent can send guidance, corrections, or new instructions based on what the subagent has reported so far
3. Use case: Main agent monitors subagent progress and provides corrective guidance during execution, rather than only being able to retry after a complete failure

Example Scenario

A coding task delegated to a subagent could:
- Report "I am about to refactor the auth module"
- Main agent responds "careful, the auth module has a pending PR, check that first"
- Subagent adjusts course mid-execution instead of failing and retrying

Suggested Approach

- Expose a callback/invoke endpoint on the main agent side that subagents can call during execution
- Or introduce a shared message bus / event queue that both main and subagent can read/write during the subagent run
- The subagent toolset would include a report_status or send_message capability that delivers to the main agent in real-time

Watch-outs

- Subagents running in isolated contexts need a transport mechanism (HTTP callback, WebSocket, etc.) to reach back to the parent
- Rate limiting / debouncing may be needed to prevent chat flooding
- This should not block or slow down the subagent primary task

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING