Skip to content

Screen mangled after some agent output #161

Description

@baruch

What happened?

I run gnhf on some task and I use claude code as the agent, the display got mangled after some time and it is now a bit off. Attached is the display, you can see the line in the middle that doesnt change and the timers do not have any text around them.

Image

I tried to refresh by the common key Ctrl-l (lower case L) and it doesnt help.

How to reproduce

No idea, it's something in my run and the outputs from claude in the repo.

gnhf version

0.1.41

Agent

claude

Model / provider (if relevant)

No response

OS and Node version

Linux Ubuntu 24.04

Debug log (gnhf.log)

{"timestamp":"2026-06-26T11:32:09.813Z","pid":495589,"event":"run:start","args":["--max-iterations","10","--current-branch","optimize the weka container startup time, to test snatch a cluster, install it with teka install it once, you can later teka upgrade it for faster turnaround. To test the container startup do  and then  on one of the hosts. You want to get the time from when you start the  until  shows the cluster is OK (fully available) as low as possible."],"runId":"optimize-the-weka-co-1c77d1","runDir":"/home/baruch/.treehouse/bare-wekapp-b749c9/3/bare-wekapp/.gnhf/runs/optimize-the-weka-co-1c77d1","agent":"claude","promptLength":342,"promptFromStdin":false,"startIteration":0,"maxIterations":10,"preventSleep":true,"worktree":false,"worktreePath":null,"currentBranch":true,"push":false,"platform":"linux","nodeVersion":"v22.22.0","gnhfVersion":"0.1.41"}
{"timestamp":"2026-06-26T11:32:09.830Z","pid":495589,"event":"orchestrator:start","agent":"claude","runId":"optimize-the-weka-co-1c77d1","startIteration":0,"maxIterations":10,"push":false,"maxConsecutiveFailures":3,"baseCommit":"ebdfae4a20a8af2f6979b32fc02dc1f6afcd5c53","initialCommitCount":0}
{"timestamp":"2026-06-26T11:32:09.840Z","pid":495589,"event":"iteration:start","iteration":1,"promptLength":2447,"consecutiveFailures":0,"totalInputTokens":0,"totalOutputTokens":0,"git":{"head":"ebdfae4a20a8af2f6979b32fc02dc1f6afcd5c53","branch":"dev/container-startup-improvement","commitCount":0}}
{"timestamp":"2026-06-26T11:32:09.841Z","pid":495589,"event":"agent:run:start","iteration":1,"agent":"claude","logPath":"/home/baruch/.treehouse/bare-wekapp-b749c9/3/bare-wekapp/.gnhf/runs/optimize-the-weka-co-1c77d1/iteration-1.jsonl"}
{"timestamp":"2026-06-26T12:02:49.198Z","pid":495589,"event":"agent:run:end","iteration":1,"elapsedMs":1839357,"success":true,"inputTokens":7257873,"outputTokens":63113,"cacheReadTokens":7230220,"cacheCreationTokens":312606,"estimated":false}
{"timestamp":"2026-06-26T12:02:49.469Z","pid":495589,"event":"iteration:end","iteration":1,"elapsedMs":1839628,"success":true,"summary":"Established the test bed and a repeatable baseline for Weka container startup: installed a persistent 5-node cluster, built a measurement harness, and measured a median 22.2s from `weka local start` to `weka status OK` with a per-phase breakdown.","keyChanges":4,"keyLearnings":5,"consecutiveFailures":0,"totalInputTokens":7257873,"totalOutputTokens":63113,"tokensEstimated":false,"commitCount":1}
{"timestamp":"2026-06-26T12:02:49.478Z","pid":495589,"event":"iteration:start","iteration":2,"promptLength":2447,"consecutiveFailures":0,"totalInputTokens":7257873,"totalOutputTokens":63113,"git":{"head":"e7ec10090c3c883aa177341604697d5e4a09801f","branch":"dev/container-startup-improvement","commitCount":1}}
{"timestamp":"2026-06-26T12:02:49.479Z","pid":495589,"event":"agent:run:start","iteration":2,"agent":"claude","logPath":"/home/baruch/.treehouse/bare-wekapp-b749c9/3/bare-wekapp/.gnhf/runs/optimize-the-weka-co-1c77d1/iteration-2.jsonl"}

notes.md (optional)

Anything else

No response

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions