claude-code - 💡(How to fix) Fix [BUG] Infrastructure Issue related to Python, Linux, Sandbox!

Preflight Checklist

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

What's Wrong?

🔴 Linux Bash Sandbox — CONFIRMED DOWN The workspace is returning: "Workspace unavailable. The isolated Linux environment failed to start." This is a hard infrastructure failure, not a configuration issue on your end. It affects every session — you can see from your session history that a recent session was literally titled "Check Python installation status," which means this has been hitting you repeatedly across multiple sessions. What's broken as a result: Any skill that runs Python scripts behind the scenes — load reviews, invoice reconciliation, settlement creation, revenue reports, xlsx/docx generation — will fail at the script execution step. The bash sandbox is what powers all of that.

✅ What's still working

File tools (Read, Write, Edit, Glob) — working normally MCP connectors (NewMile, SharePoint, Outlook) — responding fine Your outputs folder — accessible and clean All browser-based tools — working

Root cause assessment: This is on Anthropic's infrastructure side. The sandbox container for your session (blissful-practical-ramanujan) is failing to boot. You and others hitting it at the same time suggests it's either a broader outage or a persistent problem with the sandbox pool, not anything specific to your machine or account.

I have tried multiple times to restart, refresh and nothing works. Fin tells me this: When reporting, emphasize that this is a confirmed infrastructure failure affecting the sandbox container boot process, not a configuration issue, so it gets routed to the appropriate technical team for investigation.

What Should Happen?

All skills that use Python should still be working accordingly

Error Messages/Logs

I'm unable to start the SC load review — the isolated Linux workspace that runs the validation engine and report builder won't come online. I retried several times and every shell call is returning "Workspace unavailable. The isolated Linux environment failed to start."
The skill needs that environment to:

run scripts/sc_validation.py (Python + pandas) against your uploaded CSV
run scripts/build_report.js (Node + docx) to produce the Word audit report
run the W14 NewMile cross-check workflow

Without a shell, I can read the CSV via the file tools, but I can't execute the validation engine or generate the Word doc / Clean / Flagged CSV outputs that are this skill's deliverables.

Michelle — I can't run the SC load review right now. The Linux sandbox that hosts Python/pandas and Node/docx (which the validation engine and report builder need) is failing to start on this machine. I retried with several timeouts and it stays "Workspace unavailable," so I have no way to execute scripts/sc_validation.py or scripts/build_report.js, and no way to produce the Word doc and CSV outputs the skill is designed to deliver.

Steps to Reproduce

Using any skill that involves python

Claude Model

Not sure / Multiple models

Is this a regression?

Yes, this worked in a previous version

Last Working Version

opus 4.7

Claude Code Version

Cowork and Chat not code

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

PowerShell

Additional Information

Almost all of the skills I have created involve either Python, Linux, or Sandbox. Please look at the infrastructure and see what is going on.