Question 1

Does SkillzBench support Claude Code, Codex, Cursor, and Gemini CLI skills?

Accepted Answer

Yes. SkillzBench is built for SKILL.md instruction packages across Claude Code, Codex, Cursor, Gemini CLI, and chat-app agents. Launch scope covers instruction-only skills (SKILL.md plus text references); script-bearing skills are listed link-out only, so you always know what SkillzBench has verified versus what links elsewhere.

Question 2

How does SkillzBench measure whether a skill actually improves agent performance?

Accepted Answer

Every skill is benchmarked in a sandbox using the same model, tested with and without the skill, and the resulting Lift is published in percentage points. Lift is always shown separately from the trust score, and money never touches scores — so the performance number stays independent of any commercial interest.

Question 3

What does the SkillzBench security scan actually check before I install a skill?

Accepted Answer

Each skill version is pinned to a commit SHA, then validated, license-checked, security-scanned, and quality-judged. Findings feed into a 0-100 trust score built from security, quality, license, maintenance, and community signals. This addresses a real gap: published research found roughly of skills carry at least one vulnerability, and marketplace curation today is minimal.

Question 4

When can I start using SkillzBench, and how do I get early 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 as the Discover, Verify, Benchmark, Equip, and Stay Current pillars come online. Pricing is planned around a generous free tier, with Pro (~$9-15/mo), Teams (~$15-25/seat/mo), and Creator (~$5-10/mo) tiers under consideration.

Question 5

How is SkillzBench different from the existing agent skills marketplaces?

Accepted Answer

Marketplaces list and distribute skills; SkillzBench proves them. As an independent testing lab, it pins versions to commit SHAs, security-scans, license-checks, quality-judges, and benchmarks every skill with the same model — publishing Lift in percentage points that money never touches. It also goes beyond listing: one MCP server with per-agent loadouts lets you toggle skills and tools per agent, and upstream changes are tracked, re-scanned, re-benchmarked, and reported to you automatically.

Verified Agent Skills for Developers