Question 1

What exactly is an agent skill, and why does my sales team need them vetted?

Accepted Answer

Agent skills are SKILL.md instruction packages for tools like Claude Code, Codex, Cursor, Gemini CLI and chat-app agents. They're everywhere, but published research found ~carry at least one vulnerability — and malicious skills can spread via install count inflation and silent skill override. SkillzBench verifies and benchmarks each skill so you know which ones actually help before they touch your sales agents.

Question 2

How does SkillzBench measure whether a skill actually improves my AI sales agent?

Accepted Answer

We benchmark every skill in a sandbox with the same model, running the agent with and without the skill, and publish the Lift in percentage points. Lift is always shown separately from ratings, and money never touches scores — so the number reflects measured performance, not sponsorship.

Question 3

Can a skill that's safe today turn malicious after an update?

Accepted Answer

Yes — that's a documented supply chain risk: bulk update commands can turn previously benign skills malicious on the next update cycle, and skills scanned only at creation leave windows for undetected modifications. SkillzBench is building continuous protection: we track upstream changes, re-scan, re-benchmark, and notify you when something changes.

Question 4

When can my team actually start using SkillzBench loadouts?

Accepted Answer

SkillzBench is pre-launch and actively being built — the waitlist and newsletter are live now, and the product itself is not yet public. Join the waitlist to be notified at launch, when the Discover catalog, trust scores, Lift benchmarks and the Equip MCP server become available.

Question 5

How is SkillzBench different from browsing an AI agent skills marketplace directly?

Accepted Answer

Marketplaces surface skills; SkillzBench independently tests them. We pin every version to a commit SHA, security-scan and license-check each skill, publish sandbox-derived Lift numbers, and give you per-agent loadouts where the agent only sees what you chose. We're an independent testing lab and control panel — not a storefront, and money never influences scores.

Curated Agent Loadouts for Sales Teams