Question 1

How does SkillzBench measure whether an agent skill actually improves AI agent performance?

Accepted Answer

SkillzBench's Benchmark pillar tests every skill in a sandbox using the same model, running tasks both with and without the skill, then publishes the Lift in percentage points. This paired evaluation isolates the skill's actual contribution, and the Lift is always shown separately from ratings — money never touches scores.

Question 2

Can I control which skills my classroom AI assistant uses?

Accepted Answer

Yes. The Equip pillar provides one MCP server plus per-agent loadouts — you toggle skills and tools per agent, and the agent only sees what was chosen. Nothing leaks into your teaching assistant without your explicit selection.

Question 3

What about security risks in skills students might interact with?

Accepted Answer

Published research found roughly of skills carry at least one vulnerability, which is why SkillzBench's Verify pillar security-scans, license-checks, and quality-judges every skill, pins each version to a commit SHA, and assigns a 0-100 Trust Score built from security, quality, license, maintenance, and community signals.

Question 4

When will SkillzBench be available, and how do I get access?

Accepted Answer

SkillzBench is pre-launch and actively being built — the waitlist/newsletter is live now, and the product itself is not yet public. Join the waitlist to be notified at launch. Planned pricing includes a generous free tier, Pro at roughly $9-15/mo, and Teams at roughly $15-25/seat/mo with org loadouts, allowlists, and audit.

Question 5

How is SkillzBench different from just browsing skills on GitHub?

Accepted Answer

GitHub shows you a README; SkillzBench shows you evidence. Skills are continuously crawled and deduped, every version is pinned to a commit SHA and security-scanned, each skill is benchmarked with the same model with and without the skill to measure real Lift, and upstream changes are tracked with automatic re-scans and re-benchmarks. Script-bearing skills are listed link-out only during launch, keeping scope to instruction-only SKILL.md packages.

Agent Skill Benchmarking for Teachers