Question 1

How does SkillzBench handle MCP security risks like tool poisoning and prompt injection?

Accepted Answer

Every skill and server SkillzBench lists goes through security scanning as part of the Verify pillar, with each version pinned to a commit SHA. The published research found roughly of skills carry at least one vulnerability, which is exactly why verification and trust scores built from security signals exist — to surface risky packages before you enable them.

Question 2

What is the Lift score and why is it shown separately from ratings?

Accepted Answer

Lift measures what a skill actually contributes: SkillzBench benchmarks it in a sandbox with the same model, with and without the skill, and publishes the difference in percentage points. It's shown separately from trust ratings so quality and performance stay distinct, and money never touches scores.

Question 3

Can I control which MCP tools each of my agents can access?

Accepted Answer

Yes — that's the Equip pillar. You get one MCP server plus per-agent loadouts: toggle skills and tools per agent, and the agent only sees what was chosen. This directly addresses approval fatigue and excessive tool calls that slow agents down.

Question 4

When can I actually use 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 early access as the Discover, Verify, Benchmark, Equip, and Stay current pillars roll out.

Question 5

How is SkillzBench different from browsing the official MCP registry or GitHub directly?

Accepted Answer

Registries and GitHub list what exists; SkillzBench tells you what's safe and what works. Every entry is deduped, pinned to a commit SHA, license-checked, security-scanned, quality-judged, and benchmarked for Lift — with a 0-100 trust score from five signal categories and continuous upstream monitoring.

MCP Server Discovery for Designers