Claude
AI Assistant Context Leak and Hallucinated User Input Attribution
Notifications
Loading notification history…
Personal experienceImported from BlueskyPublished Oct 14, 2026
From the original report
“ME: "So, can the new Muse Wonder collect Facebook data?" AI: "You can't use Claude for that." ME: "Wait you just said Claude where did that come from?"”
Read original Bluesky ↗
Details & sources
- How received
- Imported from Bluesky
- Source published
- Oct 14, 2026
- Observed
- Unknown
- Added to CleanApp
- Oct 14, 2026
- Product connection
- Connected service
- Original issue status
- Unknown
- Resolution
- Unknown
- Verification
- Not independently reproduced
Source count unknown
Why this source type: The author shares their own direct interaction with an AI system that unexpectedly referred to Claude.
CleanApp AI summary
AI Assistant Context Leak and Hallucinated User Input Attribution: An AI assistant unexpectedly referenced Anthropic's Claude instead of maintaining its expected system context and falsely claimed the user had explicitly named Claude in the prior prompt.
Report counts are not affected-user counts. A closed original issue does not confirm a fix.
Public ID: rpt_fmuTc8lXaPG8guVD7hRKPg