benchflow-ai/benchflow

Research infra for creating RL environments, post-training, and evals.

Latest release v0.7.11 — benchflow 0.7.11 · 16 Oct 2026

10 files for Codex, Claude Code and OpenCode: benchflow-experiment-review, benchflow-traj-upload-ops, benchflow-traj-upload, benchflow and 6 more — 1,674 tokens loaded in every session.

Open on GitHub benchflow.ai

Instructions

benchflow AGENTS.md AGENTS.md A 1,663 tok
benchflow CLAUDE.md CLAUDE.md A 11 tok

Skills

benchflow-experiment-review .agents/skills/benchflow-experiment-review/SKILL.md A 99 tok
benchflow-traj-upload-ops .agents/skills/benchflow-traj-upload-ops/SKILL.md A 125 tok
benchflow-traj-upload .agents/skills/benchflow-traj-upload/SKILL.md A 98 tok
benchflow .agents/skills/benchflow/SKILL.md A 46 tok
skill-creator .agents/skills/skill-creator/SKILL.md A 64 tok
task-creator .agents/skills/task-creator/SKILL.md B 104 tok
task-review .agents/skills/task-review/SKILL.md B 141 tok
thermo-nuclear-code-quality-review .agents/skills/thermo-nuclear-code-quality-review/SKILL.md A 54 tok

These files are benchflow-ai/benchflow's own configuration — they tell Codex, Claude Code and OpenCode how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.