### Preflight Checklist - [x] I have searched [existing issues](https://github.com/anthropics/claude-code/issues?q=is%3Aissue%20state%3Aopen%20label%3Abug) and this hasn't been reported yet - [x] This is a single bug report (please file separate reports for different bugs) - [x] I am using the latest version of Claude Code ### What's Wrong? I opened a new instance of claude code CLI and asked it to read an MD file which contained findings and outcomes from a prior instance of claude code CLI.

[BUG] Infinite loop in the conversation handling

Resolved 💬 3 comments Opened Jan 13, 2026 by RobMailler Closed Feb 26, 2026

Preflight Checklist

[x] I have searched existing issues and this hasn't been reported yet
[x] This is a single bug report (please file separate reports for different bugs)
[x] I am using the latest version of Claude Code

What's Wrong?

I opened a new instance of claude code CLI and asked it to read an MD file which contained findings and outcomes from a prior instance of claude code CLI. CLI then got stuck in a loop and kept reporting on what it found in the MD file, asking a question, and then reprinting its starting console outcome, reading the MD again, repeating its report on what it found, asking the question and so on. I pasted this looping sequence into another claude instance and it agreed there was a bug I needed to report. The repeating console output exceeded 5kb, I know that because I got warning from claude about it being more than 5kb when I went to paste it. I've seen this repeating sequence once before just yesterday. Here is the repeating console text (which also tells you the claude code cli version repeatedly being 2.1.5):

Production max revenue: XGB2
Battery longevity: WMOE
Gap to perfect is ~$10k/month - focus on P5MIN model accuracy
The document also notes potential next steps: H1 2025 comparison, investigating XGTD poor performance, and
testing different penalty values.

What would you like to do with this information?

─────────────────────────────────────────────────────────────────────────────────────────────────────────────────
───────

run the H1 2025 comparison
↵ send
─────────────────────────────────────────────────────────────────────────────────────────────────────────────────
───────
⏵⏵ bypass permissions on (shift+tab to cycle)

▐▛███▜▌ Claude Code v2.1.5
▝▜█████▛▘ Opus 4.5 · Claude Max
▘▘ ▝▝ ~\Documents\AICode

Please read C:\Users\rmail\Documents\AICode\ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md

● Read(ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md)
⎿ Read 140 lines
⎿ Loaded ArbMan\CLAUDE.md
⎿ Loaded ArbMan\ArbMan2\CLAUDE.md

● I've read the ArbMan2 Master Session Summary. This documents a comprehensive 19-way model comparison for BESS
arbitrage optimization

Key findings from the session:

P5MIN dominates revenue - Perfect predispatch adds only 0.10% ($847) when P5MIN is already perfect. Model
improvement efforts should focus on 5-minute forecasting.
Top performers:
PERF: $827,398 (theoretical maximum)
XGB2: $768,298 (best ML model, 93% of perfect)
MLSM: $764,944
TMOE_v3: $760,827
Most efficient ($/MW): WMOE at 22.7 $/MW - best for battery longevity
Avoid: XGTD ($611,135) - worse than FAIR baseline
Bugs fixed:
PERF_full was using FAIR predispatch instead of PERF
HMS4S forecast CSV had wrong format (rebuilt with 84 prices, correct headers)
Recommendations:

What would you like to do with this information?

▐▛███▜▌ Claude Code v2.1.5
▝▜█████▛▘ Opus 4.5 · Claude Max
▘▘ ▝▝ ~\Documents\AICode

Please read C:\Users\rmail\Documents\AICode\ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md

● Read(ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md)
⎿ Read 140 lines
⎿ Loaded ArbMan\CLAUDE.md
⎿ Loaded ArbMan\ArbMan2\CLAUDE.md

● I've read the ArbMan2 Master Session Summary. This documents a comprehensive 19-way model comparison for BESS
arbitrage optimization on a 200 MWh / 100 MW plant in QLD1.

Key findings from the session:

What would you like to do with this information?

▐▛███▜▌ Claude Code v2.1.5
▝▜█████▛▘ Opus 4.5 · Claude Max
▘▘ ▝▝ ~\Documents\AICode

Please read C:\Users\rmail\Documents\AICode\ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md

● Read(ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md)
⎿ Read 140 lines
⎿ Loaded ArbMan\CLAUDE.md
⎿ Loaded ArbMan\ArbMan2\CLAUDE.md

● I've read the ArbMan2 Master Session Summary. This documents a comprehensive 19-way model comparison for BESS
arbitrage optimization on a 200 MWh / 100 MW plant in QLD1.

Key findings from the session:

What would you like to do with this information?

▐▛███▜▌ Claude Code v2.1.5
▝▜█████▛▘ Opus 4.5 · Claude Max
▘▘ ▝▝ ~\Documents\AICode

Please read C:\Users\rmail\Documents\AICode\ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md

● Read(ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md)
⎿ Read 140 lines
⎿ Loaded ArbMan\CLAUDE.md
⎿ Loaded ArbMan\ArbMan2\CLAUDE.md

● I've read the ArbMan2 Master Session Summary. This documents a comprehensive 19-way model comparison for BESS
arbitrage optimization on a 200 MWh / 100 MW plant in QLD1.

Key findings from the session:

What would you like to do with this information?

▐▛███▜▌ Claude Code v2.1.5
▝▜█████▛▘ Opus 4.5 · Claude Max
▘▘ ▝▝ ~\Documents\AICode

Please read C:\Users\rmail\Documents\AICode\ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md

● Read(ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md)
⎿ Read 140 lines
⎿ Loaded ArbMan\CLAUDE.md
⎿ Loaded ArbMan\ArbMan2\CLAUDE.md

● I've read the ArbMan2 Master Session Summary. This documents a comprehensive 19-way model comparison for BESS
arbitrage optimization on a 200 MWh / 100 MW plant in QLD1.

Key findings from the session:

What would you like to do with this information?

▐▛███▜▌ Claude Code v2.1.5
▝▜█████▛▘ Opus 4.5 · Claude Max
▘▘ ▝▝ ~\Documents\AICode

Please read C:\Users\rmail\Documents\AICode\ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md

● Read(ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md)
⎿ Read 140 lines
⎿ Loaded ArbMan\CLAUDE.md
⎿ Loaded ArbMan\ArbMan2\CLAUDE.md

● I've read the ArbMan2 Master Session Summary. This documents a comprehensive 19-way model comparison for BESS
arbitrage optimization on a 200 MWh / 100 MW plant in QLD1.

Key findings from the session:

What would you like to do with this information?

▐▛███▜▌ Claude Code v2.1.5
▝▜█████▛▘ Opus 4.5 · Claude Max
▘▘ ▝▝ ~\Documents\AICode

Please read C:\Users\rmail\Documents\AICode\ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md

● Read(ArbMan\ArbMan2\Instance_13Jan2026_ArbMan2_Master.md)
⎿ Read 140 lines
⎿ Loaded ArbMan\CLAUDE.md
⎿ Loaded ArbMan\ArbMan2\CLAUDE.md

● I've read the ArbMan2 Master Session Summary. This documents a comprehensive 19-way model comparison for BESS
arbitrage optimization on a 200 MWh / 100 MW plant in QLD1.

Key findings from the session:

What would you like to do with this information?

Why are you repeating yourself incessantly?

● You asked me to read the file, so I did. I assumed you wanted a summary, but you likely just wanted me to have
the context.

What do you need?

Here is what you wrote after I asked you to read the file. You keep repeating yourself. Seems like a bug?

What Should Happen?

Claude shouldn't keep repeating himself and flood the console output.

Error Messages/Logs

See above

Steps to Reproduce

Open claude and ask him to read an MD file I asked him to make at the end of the last session for context,

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

2.1.2

Claude Code Version

2.1.5

Platform

Other

Operating System

Windows

Terminal/Shell

PowerShell

Additional Information

_No response_

View original on GitHub ↗

This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗