[BUG] PDF API Error

Environment

• Platform (select one):
• [x ] Anthropic API
• [ ] AWS Bedrock
• [ ] Google Vertex AI
• [ ] Other: <!-- specify -->
• Claude CLI version: <!-- output of claude --version -->
• Operating System: <!-- Windows 11, Ubuntu 22.04 -->
• Terminal: <!-- Ubuntu, PowerShell -->

Bug Description

<!-- A persistent API error occurs when trying to read a .pdf file that's too big, forcing the user to close and re-open Claude Code -->

Steps to Reproduce

1. <!-- Tell Claude to read a .pdf that's, say, 155 pages -->
2. <!-- You receive an error message like:

"API Error: 400 {"type":"error","error" {"type":"invalid_request_error","message":"messages.12.content.0.pdf.source.base64.data: A maximum of 100 PDF pages may be provided."},"request_id":"req_011CSKDpX5YGUmAnT67cA9qx"}" -->

1. <!-- No matter what you try to do afterward, the error repeats and repeats until you close the app -->

Expected Behavior

<!-- If there's a limit for the number of pages that Claude can read, I would expect Claude to either break it up into sections for each session or simply inform the user that the document is too large and perhaps suggest breaking the document up into smaller documents -->

Actual Behavior

<!-- API Error: 400 {"type":"error","error":{"type":"invalid_request_error","message":"messages.12.content.0.pdf.source.base64.data: A maximum of 100 PDF pages may be provided."},"request_id":"req_011CSKDpX5YGUmAnT67cA9qx"} -->

Additional Context

<!-- Add any other context about the problem here, such as screenshots,

<img width="3612" height="1615" alt="Image" src="https://github.com/user-attachments/assets/400748f2-dbc9-49d1-8c4f-9990646398be" />

logs, etc. -->