Claude Projects isolation appears broken: model recalls content from other projects
MaterialDurian2836 · reddit · 2026-09-12
A Reddit user reports that inside one Claude Project, the model repeatedly referenced content that existed only in other Projects — including project-specific terms — and even named those projects when asked, despite the user never attaching that information to the current chat.
The user valued the per-project memory isolation and wants to know whether this is a new feature or a bug, and how to revert. There is no official response yet. If confirmed, this is a notable privacy concern for anyone using Projects to keep client or sensitive material compartmentalized.
More from Models
- Anthropic removes 5-hour usage cap for verified Claude Ultra subscribers — james_mtc · 2026-10-31
- DeepSeek's silent downtime upgrade caught by the pelican bicycle test — teortaxesTex · 2026-10-31
- ARC Prize 2026 closes in 3 days as Yi-Chia Chen tops ARC-AGI-3 at 75.56% — GregKamradt · 2026-10-31
- Fable 5.5 Max stuns with gorgeous SVG of a pelican riding a bicycle, headed to Claude Code — ChrisGPT · 2026-10-31
- OpenAI reportedly to train on users' Daybreak sessions with no opt-out — dyn___ · 2026-10-31
- NVIDIA Nemotron 3.5 scales LLM self-verification, +9-10 points on agentic benchmarks — sarahcat21 · 2026-10-31