README.md

metadata

title: README
emoji: 😻
colorFrom: purple
colorTo: purple
sdk: static
pinned: false

Train single and multi agents to rebuild programs with RL

1,261 environments where an agent runs a compiled program it cannot read, then rebuilds the whole program. Hidden behavioral tests score the result.

• ProgramLab: tasks, metadata, and build files
• ProgramLab-Tests: hidden tests
• Prebuilt task images: ghcr.io/programlab-agent/programlab