Question 1

How does SkillzBench verify a SKILL.md skill I've published on GitHub?

Accepted Answer

SkillzBench continuously crawls and dedupes skills from GitHub. When you join as a verified author, every version of your skill is pinned to a commit SHA, then validated, license-checked, security-scanned, and quality-judged by the independent lab before earning verified status.

Question 2

What exactly is the Lift score and how is it measured for my skill?

Accepted Answer

Lift is measured by running your skill in a sandbox with the same model twice — once with the skill and once without — and publishing the difference in percentage points. It is always shown separately from ratings, and money never touches scores, so your Lift is a purely objective measure.

Question 3

Does SkillzBench support script-bearing skills, or only instruction-only SKILL.md packages?

Accepted Answer

At launch, verification covers instruction-only skills — SKILL.md plus text references. Script-bearing skills are listed link-out only, so authors of script-based packages can still be discovered but won't receive full sandbox benchmarking initially.

Question 4

What happens when I update my skill after it's verified?

Accepted Answer

SkillzBench's Stay current pillar tracks upstream changes to your skill, re-scans it for security issues, re-benchmarks it, and notifies you of the results. Your verification stays fresh instead of going stale as you ship updates.

Question 5

How is SkillzBench different from just listing my skill on GitHub with a good README?

Accepted Answer

A README is a self-published claim; SkillzBench is an independent lab. You get commit-SHA-pinned versions, security scans, license checks, quality judging, a same-model Lift benchmark published in percentage points, and a 0-100 trust score from security, quality, license, maintenance and community signals — none of which a README can credibly provide.

Skill Author Verification for Writers