Question 1

What exactly does SkillzBench benchmark, and how is the Lift calculated?

Accepted Answer

SkillzBench tests instruction-only skills (SKILL.md plus text references) in a sandbox using the same model, running tasks with and without the skill. The difference in performance is published as the Lift in percentage points, always shown separately from ratings so you can judge capability and quality independently.

Question 2

How does SkillzBench protect my sales agents from vulnerable skills?

Accepted Answer

Every skill version is pinned to a commit SHA and passes validation, license checks, security scans, and quality judging. You also get a 0-100 trust score built from security, quality, license, maintenance, and community signals — and money never touches scores.

Question 3

Can I control which skills each of my agents can see?

Accepted Answer

Yes. SkillzBench is building Equip: one MCP server plus per-agent loadouts. You toggle skills and tools per agent, and the agent only sees what was chosen — no accidental exposure of the full skill library.

Question 4

When can I actually start using SkillzBench?

Accepted Answer

SkillzBench is pre-launch and actively being built — the waitlist and newsletter are live now, but the product itself is not yet public. Joining the waitlist gets you notified first when benchmarking, scanning, and the MCP server become available. Launch scope covers instruction-only skills; script-bearing skills are listed link-out only.

Question 5

How is SkillzBench different from public agent leaderboards like SWE-bench or tau-bench rankings?

Accepted Answer

Public leaderboards score agents, often with single runs vulnerable to seed noise and configuration swings. SkillzBench scores skills, not agents — in a controlled sandbox with the same model, with and without the skill — and pairs the published Lift with a trust score from security, quality, license, maintenance, and community signals. It's a control panel for what goes into your agents, not a vanity ranking.

Agent Skill Benchmarking for Sales Teams