This PR adds evaluation results in the [`.eval_results/` format](https://huggingface.co/docs/hub/main/en/eval-results#adding-evaluation-results), so they show up on the official benchmark leaderboards on the Hub.
- **Terminal-Bench-Science 0.1** (`harborframework/terminal-bench-science`, `terminalbenchscience_0_1`): 2.4, from the [Artificial Analysis leaderboard](https://artificialanalysis.ai/evaluations/terminal-bench-science) (Step 5 Preview).
Artificial Analysis runs these benchmarks independently with the mini-swe-agent harness and reports pass@1 averaged over three repeats per task.
