Verified scores

What it is

A verified score is a numeric rating attached to a (code agent, model, operation) triple — for example “claudecode/opus4_7 / implementation” or “geminicli/gemini-2.5-pro / code-review”. Every time a user completes a skill they are prompted for a 1-5 satisfaction rating; those ratings are stored, time-windowed (all-time / month / week), and aggregated into score buckets:

Scores are surfaced in the Settings TUI and on the model entry pages so you can see which agent/model combinations are reliable for which kinds of work in your project.

Why it exists

The pace of new model releases makes any static “best model” recommendation immediately stale. Verified scores let your own usage history drive that recommendation: each project accumulates evidence about which models reliably plan, implement, and review your kind of code, and the framework surfaces that evidence at the points where you choose a model.

How to use

The complete schema, score derivation, and CLI tooling are documented on the Verified scores skill page.

See also

• Agent attribution — how a task records the model that implemented it
• Settings TUI — where scores are displayed

Next: Agent attribution