Research

Researching how agents learn, act, collaborate, and improve.

Aug 11, 2026

VibeLifeBench: Can Your Life Agent Be Proactive and Persistent in a Living World

Personal assistants must operate in a world that keeps changing, often for weeks and without continuous user prompts. We introduce VibeLifeBench, a benchmark of 200 long-horizon tasks across ten everyday-life domains and 22 simulated services. Seven frontier models all perform poorly, revealing a fundamental gap between agents that respond to requests and agents that can proactively monitor change, preserve intent, and act consistently over time.

Jul 28, 2026

RSIBench-Data: Benchmarking Data-Centric Research for Recursive Self-Improvement

Can agents turn evidence of model failures into better training data and improved models? We introduce RSIBench-Data, a controlled benchmark that isolates data-centric research decisions within a fixed post-training stack. Across six benchmarks, agents show that they can form useful hypotheses and improve initial attempts, but struggle to preserve strong results and convert feedback into consistent progress—highlighting a central bottleneck in recursive self-improvement.

May 15, 2026

The Scaling Laws of Skills in LLM Agent Systems

What happens when an agent’s reusable skill library grows from dozens of skills to hundreds or thousands? Across 15 frontier models, 1,141 real-world skills, and more than three million routing and execution decisions, we identify coupled scaling laws governing skill selection and task execution. The results show that agent performance depends not only on model capability, but also on how skills are structured, differentiated, and exposed.

May 14, 2026

Do Coding Agents Understand Least-Privilege Authorization?

Coding agents increasingly operate on shells, repositories, and user files, making least-privilege authorization essential for safe deployment. We introduce AuthBench, a benchmark of 120 realistic terminal tasks with human-reviewed permission boundaries and executable security checks. Current models routinely grant unnecessary access while omitting permissions required for execution, showing that authorization cannot be solved by simply making agents more conservative or giving them more reasoning time.

Apr 26, 2026

ClawMark: A Living-World Benchmark for Multi-Turn, Multi-Day, Multimodal Coworker Agents

Coworker agents must complete workflows that span multiple days while emails, calendars, documents, and multimodal evidence continue to change around them. We introduce ClawMark, a benchmark of 100 tasks across 13 professional scenarios, evaluated through stateful sandboxed services and deterministic rule-based checks. Frontier agents often make partial progress but rarely complete the full workflow, with performance dropping sharply after the environment changes.

VibeLifeBench: Can Your Life Agent Be Proactive and Persistent in a Living World

RSIBench-Data: Benchmarking Data-Centric Research for Recursive Self-Improvement

The Scaling Laws of Skills in LLM Agent Systems

Do Coding Agents Understand Least-Privilege Authorization?

ClawMark: A Living-World Benchmark for Multi-Turn, Multi-Day, Multimodal Coworker Agents