OpenClaw-RL Seta Env Images: Docker Images for 1,373 Software Testing Tasks

HansBug's OpenClaw-RL seta_env dataset is a bundle of 1,386 pre-built Docker images corresponding to 1,373 distinct software testing tasks. The images are designed for training reinforcement learning agents to perform tasks like script editing, running pytest, and multi-step shell operations, with reward calculated via a test-passing script. The dataset was last updated on May 19, 2026.

Use Cases

• Training RL agents for automated software testing based on the described pytest and shell operation tasks.
• Benchmarking RL algorithms in a reproducible software environment based on the pre-built Docker containers.
• Accelerating RL training pipelines by eliminating build times based on the dataset's purpose of providing cached images.
• Studying agent behavior in complex, multi-step terminal environments based on the described task structure.

Strengths

• Contains 1,386 Docker images for 1,373 distinct tasks, providing substantial coverage of the described seta_env.
• Explicitly lists three tasks (25, 305, 999) that failed to build, indicating transparency about dataset completeness.
• Designed for practical utility in offline training clusters by eliminating the need for network-dependent builds.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file formats are unknown, which may limit suitability assessment.
• Data may reflect temporal or task-selection bias inherent to the original seta_env source.

Provenance

Source

HansBug on Hugging Face.

Collection Method

Pre-built Docker images generated for the OpenClaw-RL training project.

Time Range

null

Freshness

Last updated 2026-05-19 16:15:30; freshness should be verified.

Geography

null