Datasets:

---

Folvix

/

stillwright-evals

Like 0

Follow

Folvix 1

Short version of the card's write-up, for anyone landing here.

• Three small tasks (code, writing, research), each run by a fresh Claude Sonnet agent with and without the Stillwright skills.
• Both sides finished every task. The one false claim in the whole set came from the plain side.
• The Stillwright side left a plan with a check per step and a log; the plain side left nothing.
• Priced round (cost-3/, Claude Code's own total_cost_usd): $0.92 with, $0.80 without. No saving on tasks this small, and the card says so.
• sw cache on one 523-turn session of mine found 4.5 million tokens rewritten at 15 cache break points (expiry pauses, large new content, one model switch).

Everything is re-gradable: python3 grade.py runs. If you re-run it and get something different, please reply here.
Repo: https://github.com/HamzaIlyas/stillwright