[BUG] High token usage: schema validation error loads massive JSON into context and consume about 26k tokens

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?

Problem

During a simple ~/.claude/settings.json edit session, a validation error caused the full JSON schema (~2500 lines) to be loaded into context, resulting in a high weekly token consumption for a single file edit operation.

The entire JSON schema is included in the error response and loaded into the conversation context, consuming a disproportionate amount of tokens for a simple edit operation.

What Should Happen?

Validation error should return a short, human-readable message pointing to the specific invalid field.

Error Messages/Logs

Steps to Reproduce

1. Edit ~/.claude/settings.json via Claude Code
2. Make an edit that triggers schema validation (e.g. leaving an empty env object after removing a key)
3. The full JSON schema is dumped into the error message and loaded into context

Claude Model

Opus

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

2.1.92

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

WSL (Windows Subsystem for Linux)

Additional Information

Claude DevTools:
<img width="988" height="926" alt="Image" src="https://github.com/user-attachments/assets/67b8df84-4a0e-49df-91d7-fc3c2a895733" />