Scale Labs Open-Sources AgentEnv for Modular Multi-Agent RL Environments

Scale Labs open-sources AgentEnv, a modular multi-agent reinforcement-learning framework that runs the same task across Claude Code, OpenAI, and custom agents using Environment and Agent Cards, with 48 built-in step types and four sandbox options.

• Scale Labs open-sources AgentEnv, a framework for building modular RL environments that supports multiple tasks and agents.
• AgentEnv uses Environment Cards and Agent Cards to let the same task run without custom wiring on Claude Code, OpenAI agents, or custom agents via MCP, REST, generated CLI, and A2A.
• The framework ships with 48 built-in step types and four sandbox options: local Docker, Modal containers, Modal virtual machines, and E2B virtual machines.