Issue Station
Sign In
Tools hereToolbox

hermes - ✅(Solved) Fix [Bug]: Context bursting in long context session [1 pull requests, 3 comments, 3 participants]

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…
GitHub stats
NousResearch/hermes-agent#11529•Fetched 2026-04-18 06:00:32
View on GitHub
Comments
3
Participants
3
Timeline
6
Reactions
0
Author
0xharryriddle0xharryriddle
Participants
0xharryriddle0xharryriddlegongli0929gongli0929ygd58ygd58
Timeline (top)
commented ×3cross-referenced ×1labeled ×1referenced ×1

Error Message

Additional Logs / Traceback (optional)

Root Cause

Root Cause Analysis (optional)

Fix Action

Fixed

PR fix notes

PR #11559: fix(compressor): skip compression during summary LLM cooldown to prevent CLI freeze

Description (problem / solution / changelog)

Problem

When the summary LLM returns 429 or 401, _generate_summary() sets a cooldown and returns None. compress() inserts a static fallback marker and returns normally. However should_compress() did not check the cooldown, so every subsequent agent turn triggered _compress_context() again — the fallback marker was inserted repeatedly, the loop never exited, and the CLI appeared frozen.

From the reporter log:

ERROR: API call failed after 3 retries. HTTP 429: ... msgs=455 tokens=~327,041
patch 60.3s [error]

Fix

Add a cooldown guard in should_compress(): if _summary_failure_cooldown_until is in the future, return False. Reuses the existing float that _generate_summary() already maintains — no new state needed.

Result

During cooldown the agent loop continues normally and the CLI remains responsive. When the cooldown expires, compression is retried automatically.

Fixes #11529

Changed files

  • agent/context_compressor.py (modified, +14/-1)

Code Example

Debug report uploaded:
  Report     https://paste.rs/Umtz3
  agent.log  https://paste.rs/jRWGA

---
RAW_BUFFERClick to expand / collapse

Bug Description

I'm stucking in here and cannot press anything or type any message in the long context. In addition, I have updated the "Max turns" of model to 999 for avoid facing max interaction issue, then this issue was appearing

<img width="1283" height="208" alt="Image" src="https://github.com/user-attachments/assets/741a8578-ccb1-4da3-a6bd-a8dbbb4501e1" />

Steps to Reproduce

  1. Run hermes update
  2. Run hermes --yolo

Expected Behavior

Continue working on the session

Actual Behavior

Stucking in here and did not show anything

Affected Component

CLI (interactive chat)

Messaging Platform (if gateway-related)

No response

Debug Report

Debug report uploaded:
  Report     https://paste.rs/Umtz3
  agent.log  https://paste.rs/jRWGA

Operating System

Ubuntu 24.04

Python Version

3.11.14

Hermes Version

0.10.0

Additional Logs / Traceback (optional)

Root Cause Analysis (optional)

No response

Proposed Fix (optional)

No response

Are you willing to submit a PR for this?

  • I'd like to fix this myself and submit a PR

extent analysis

TL;DR

The issue may be related to the recent update of "Max turns" to 999, causing the interactive chat to freeze, and checking the debug report could provide more insights.

Guidance

  • Review the debug report at https://paste.rs/Umtz3 and agent.log at https://paste.rs/jRWGA to identify any error messages or clues that could point to the cause of the issue.
  • Try resetting the "Max turns" to its original value to see if the issue persists, which could help determine if the update is indeed the cause.
  • Run hermes --yolo with the --verbose flag (if available) to get more detailed output that might help diagnose the problem.
  • Check if there are any known issues or updates related to Hermes version 0.10.0 that could be contributing to this behavior.

Example

No specific code snippet can be provided without more context, but reviewing the debug reports and logs is a crucial step.

Notes

The issue seems to be related to the interactive chat feature of the Hermes CLI, and the provided debug reports and logs should be carefully examined to understand the root cause. Without more information or context about the Hermes application and its configuration, it's challenging to provide a definitive fix.

Recommendation

Apply workaround: Try resetting the "Max turns" value and review the debug reports to gather more information before attempting any code changes or updates, as this approach allows for a safer and more informed next step.

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…
#file not found#serialization error#model compatibility#GPU setup#container setup

Still need to ship something?

×6

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

Back to top recommendations

TRENDING

hermes - ✅(Solved) Fix [Bug]: Context bursting in long context session [1 pull requests, 3 comments, 3 participants]