How we score

Every skill gets two scores. We never mix them.

Critic Score

Dial: DangerDanger0–39
Dial: WeakWeak40–59
Dial: GoodGood60–84
Dial: ExcellentExcellent85–100

From our own tests, run the same way every time, pinned to the skill version and the model. Anyone can rerun them.

Crowd Score

4.6/ 5
  • Only signed-in users can rate
  • One rating per person
  • Shown next to the Critic Score, never added to it

What our tests check

Starts at the right timeDoes the skill switch on when it should, and stay quiet when it should not?
Gets the job doneDoes the agent do the task better with the skill than without it?
Safe to installNo hidden code, data leaks or tricks in the instructions.
Easy to followAre the instructions clear and short enough for an agent to use?
Allowed to useIs the licence clear and does the author check out?

How it looks on a skill

Dial: Good

Critic Score 82 / 100

Good: our reproducible tests.

4.6

Crowd Score 4.6 / 5

128 reviews from people who use it

Scores agree

Test results

Starts at the right time88%
Gets the job done76%
Safe to install100%
Easy to follow82%
Allowed to use92%

Example values for illustration.

Where the list comes from

The library starts from the most-installed skills on public sources (skills.sh, ClawHub and Anthropic's official skills) plus submissions a person has checked. Their rankings and counts are shown with attribution and a date. Popularity is not a review.

Authors

We list facts and links only; we do not host or copy skill files. Authors can ask for corrections or removal at any time.