Who · How · Why · Edu Tor

Agent-side methodology

Agent Twin Scores are editorial 0–100 composites that weight API quality, MCP or tool surface, CLI, structured docs, machine auth, quotas, determinism, and skill packaging. Named staff assign them from public products and public documentation — not from invented user counts and not from a 3D canvas.

Who, how, and why

Who: Edu Tor (editor), with staff writers Maya Chen and Jonas Okonkwo on the people twin. How: public apps, official docs, pricing pages, and capability matrices in the committed catalog. Why: a citable machine contract, not search bait.

SKILL.md and the dataset

Each product has a real SKILL.md derived from catalog facts, with explicit when-to-use and when-not-to-use. Download JSON/CSV under CC BY 4.0 from /dataset/twin-scores.json and recompute. Rubric: scorecard.