Question 1

How does SkillzBench verify a SKILL.md skill differently from a registry listing?

Accepted Answer

Registries list skills; SkillzBench verifies them. Every version is pinned to a commit SHA, then validated, license-checked, security-scanned, and quality-judged. The result is a 0-100 trust score from security, quality, license, maintenance, and community signals — plus a published Lift from sandbox benchmarking. Launch scope is instruction-only skills (SKILL.md plus text references); script-bearing skills are listed link-out only.

Question 2

Can I pay for a higher trust score or better benchmark results?

Accepted Answer

No. Money never touches scores — that's a core principle. The Creator tier (~$5-10/mo) gets you verified-author status, and benchmark-as-a-service is available for priority runs, but scores are computed independently from security, quality, license, maintenance, and community signals. Lift is always shown separately from ratings so nobody can game the system.

Question 3

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

Accepted Answer

The stay-current pillar tracks upstream changes, re-scans, re-benchmarks, and notifies you. Each new version gets pinned to its own commit SHA and goes through the same verification pipeline, so your published trust score and Lift stay accurate over time rather than going stale.

Question 4

How does the benchmark actually measure whether my skill helps agents?

Accepted Answer

SkillzBench's Benchmark pillar tests every skill in a sandbox with the same model, running tasks both with and without the skill, then publishes the Lift in percentage points. Because the model and conditions are held constant, the Lift is a like-for-like comparison buyers can trust — not a self-reported claim.

Question 5

Why should I trust SkillzBench instead of the vetting on existing agent skill marketplaces?

Accepted Answer

SkillzBench is an independent testing lab run by AIQ Labs — not a marketplace that profits from listings. Research found ~26% of skills carry at least one vulnerability, and major registries either do no security vetting or score quality without checking security. SkillzBench separates those concerns: security scanning, license checks, quality judging, and benchmarking each feed an independent trust score, and money never touches scores.

Skill Author Verification for Teachers