Decision information
Verified: 8/14/2026- Ecosystem category
- Claw Agents
- Agent class
- Research
- Lineage
- Research
- Runs
- Local
- Autonomy
- Autonomous
- Provider
- Claw AI Lab
- Content type
- Product
- Pricing
- Open Source
- Open source
- Yes
- License
- Not recorded
- Maintenance
- Active
- Install methods
- Source
- Platforms
- Linux, Local GPU
- Interfaces
- Web, Terminal
Documented Claw signals
View inclusion standard →Tool useMulti-step goalsProactive communicationAutonomous executionSkills / plugins / MCPPersistent / long-runningMemory / stateBrowser / shell / computer access
Security signals (informational, not a security certification)
Source availableRequires shellRequires credentials
Verification source: https://arxiv.org/abs/2605.22662 · preview-v1.1.0 / arXiv:2605.22662
Overview
A lab-native autonomous multi-agent research team with configurable roles, collaborative workflows, a PI dashboard, experiment execution, monitoring, and rollback or resume controls.
What it provides
- Creates a configurable research team from one prompt
- Supports exploration, multi-agent discussion, and reproduction modes
- Connects code, datasets, checkpoints, and experiment artifacts through Claw-Code Harness
AllClaw classification
This project is listed as a Claw because its documented behavior includes several category signals such as persistence, autonomous execution, tools, memory, multi-step work, scheduling, proactive communication, or extensibility.
Provider: Claw AI Lab. Primary source: https://arxiv.org/abs/2605.22662.
The classification and core claims were checked against the linked primary source on August 14, 2026.
Tags
research-agentmulti-agentexperimentsdashboardautonomous-research