A benchmark of data-analysis traps for coding agents

This benchmark uncovers data‑analysis pitfalls for coding agents, enabling quick assessment of their weaknesses. To adopt it, integrate the repo into your local