[MODEL] Claude continues to fabricate information
Preflight Checklist
- [x] I have searched existing issues for similar behavior reports
- [x] This report does NOT contain sensitive information (API keys, passwords, etc.)
Type of Behavior Issue
Claude modified files I didn't ask it to modify
What You Asked Claude to Do
I asked Claude to confirm what notifications we send out for a given feature. I usually ask it to confirm in certain code bases and tools, and also flag that it should surface any assumptions its making. In spite of this, Claude continues to report inaccurate information and present it as validated. e.g. it said certain reminders are sent on Day X, but that wasn't true. I pushed back on Claude, and it admitted that it had fabricated that information.
I cannot trust a model that continues to make things up.
What Claude Actually Did
I asked Claude to confirm what notifications we send out for a given feature. I usually ask it to confirm in certain code bases and tools, and also flag that it should surface any assumptions its making. In spite of this, Claude continues to report inaccurate information and present it as validated. e.g. it said certain reminders are sent on Day X, but that wasn't true. I pushed back on Claude, and it admitted that it had fabricated that information.
I cannot trust a model that continues to make things up.
Expected Behavior
Stop lying.
Files Affected
Permission Mode
Accept Edits was ON (auto-accepting changes)
Can You Reproduce This?
Yes, every time with the same prompt
Steps to Reproduce
I have tried with similar prompts, and Claude continues to fabricate details.
Claude Model
Opus
Relevant Conversation
Impact
Medium - Extra work to undo changes
Claude Code Version
claude-opus-4-6
Platform
Anthropic API
Additional Context
_No response_