Question 1

How does SkillzBench verify a SKILL.md skill before I use it in Claude Code?

Accepted Answer

SkillzBench's Verify pillar pins every skill version to a commit SHA, then validates the instruction package, checks its license, runs security scans, and applies quality judging. Instruction-only skills (SKILL.md plus text references) get full treatment; script-bearing skills are listed link-out only at launch, so you always know the verification depth before equipping anything.

Question 2

What does 'Lift in percentage points' mean when testing agent skills?

Accepted Answer

SkillzBench benchmarks every skill in a sandbox using the same model, running tasks both with and without the skill installed. The Lift is the difference in performance expressed in percentage points, and it's always shown separately from the trust score and ratings — so a well-marketed skill can't hide behind a good logo or a high rating.

Question 3

Is SkillzBench free for students, and what do paid tiers cost?

Accepted Answer

SkillzBench is planning a generous free tier, with a Pro tier at roughly $9-15/month for more agents and unlimited enabled skills. A Creator tier (~$5-10/month) will serve verified authors. Exact pricing will be finalized at launch — join the waitlist to be notified first.

Question 4

When can I actually use SkillzBench, and how do I get early access?

Accepted Answer

SkillzBench is pre-launch and actively being built — the waitlist and newsletter are live now, while the product itself is not yet public. Joining the waitlist gets you launch announcements and early access as the Discover, Verify, Benchmark, Equip, and Stay Current pillars come online.

Question 5

How is SkillzBench different from just browsing skills on GitHub or a community skill registry?

Accepted Answer

GitHub shows you a README, not evidence. SkillzBench continuously crawls and dedupes skills, pins versions to commit SHAs, security-scans and license-checks each one, benchmarks them in a sandbox with published Lift numbers, and assigns a 0-100 trust score from security, quality, license, maintenance, and community signals. Money never touches scores, so rankings stay independent — and per-agent loadouts via one MCP server mean your agent only sees vetted skills you chose.

Skill Author Verification for Students