[Bug] Text truncation and character corruption when pasting from documents

Resolved 💬 3 comments Opened Mar 1, 2026 by darianrosebrook Closed Mar 5, 2026

Bug Description
Hey, often when pasting text from any document (vs code, cursor, the web, a finder preview snippet, I'm getting odd truncation of words that I worry is gumming up the context I'm trying to provide claude. I've pasted this paragraph:

D3-B scope guard (explicitly not required yet)

No multi-pack transfer with baseline proposer on the first pass. Keep pack fixed so failures are attributable to the integration chain (linearizer → proposer → seam) rather than to domain differences.

No requirement that passed_strict be true. You’re proving seam coherence, not semantic policy convergence.

which ends up as

  D3-B scope guard (explicitly not required yet)                                                                                                                        
                                                                                                                                                                        
  * Nck transfer with baseline proposer on the first pass. Keep pack fixed so failures are attributable to the integration chain (linearizer → proposer → seam) rather  
  than to domain differences.                                                                                                                                           
  * No requirement that `passed_strict` be true. You’re proving seam coherence, not semantic policy convergence.                    

I'm worried that important instructions may get truncated or missed entirely because the paste turns some of the context to gibberish.

I've also confirmed this is the text received by looking at the content in ~/.claude

Environment Info

  • Platform: darwin
  • Terminal: cursor
  • Version: 2.1.63
  • Feedback ID: 2e0f233b-fa17-4d5f-a23c-c49f105f98e4

Errors

[{"error":"Error: LSP server plugin:rust-analyzer-lsp:rust-analyzer crashed with exit code 1\n    at <anonymous> (/$bunfs/root/claude:2716:21470)\n    at emit (node:events:98:22)\n    at #handleOnExit (node:child_process:520:14)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-03-01T04:04:55.528Z"},{"error":"MaxFileReadTokenExceededError: File content (47064 tokens) exceeds maximum allowed tokens (25000). Please use offset and limit parameters to read specific portions of the file, or use the GrepTool to search for specific content.\n    at URB (/$bunfs/root/claude:1766:7815)\n    at async fRB (/$bunfs/root/claude:1770:2491)\n    at async call (/$bunfs/root/claude:1775:6062)\n    at async BZ8 (/$bunfs/root/claude:3507:15122)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-03-01T05:08:00.882Z"},{"error":"Error: EISDIR: illegal operation on a directory, read '/Users/darianrosebrook/Desktop/Projects/sterling/core/reasoning/loop'\n    at le_ (/$bunfs/root/claude:1652:420)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-03-01T05:09:26.305Z"}]

View original on GitHub ↗

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