ATANT: An Evaluation Framework for AI Continuity

Create an evaluation framework that measures 'continuity' in AI systems over time. Go beyond static benchmarks and test how well an agent maintains context across different sessions and memory stores.

Suggested repo: ContinuityBench

"Does your agent really remember? Benchmark true context-persistence."

Estimated effort: 35h