Question 1

How does SkillzBench security-scan an agent skill?

Accepted Answer

SkillzBench's Verify pillar pins every skill version to a commit SHA, validates the instructions, runs a license check, performs a security scan, and applies a quality judgment. Results feed into a 0-100 trust score built from security, quality, license, maintenance, and community signals — and the benchmark Lift is always published separately from that rating.

Question 2

Can a malicious SKILL.md package compromise my Claude Code or Cursor agent?

Accepted Answer

Instruction packages are powerful because agents follow them literally, and published research found ~of skills carry at least one vulnerability. SkillzBench is building the safety layer: continuous crawling and deduping of skills from GitHub, security scanning before you install, and per-agent loadouts so an agent only ever sees the skills and tools you explicitly enabled.

Question 3

Does SkillzBench support script-bearing skills and custom tools?

Accepted Answer

At launch, SkillzBench covers instruction-only skills — SKILL.md plus text references. Script-bearing skills are listed link-out only, so you know they exist but they do not run through the platform's verified pipeline. This keeps the launch scope tight and the scanning trustworthy.

Question 4

When can I start using SkillzBench, and what will it cost?

Accepted Answer

SkillzBench is pre-launch and actively being built — the waitlist and newsletter are live now. The pricing hypothesis includes a generous free tier, Pro around $9-15/month for more agents and unlimited enabled skills, Teams around $15-25/seat/month for org loadouts, allowlists, and audit, and Creator around $5-10/month for verified authors.

Question 5

How is SkillzBench different from just browsing GitHub for skills?

Accepted Answer

GitHub gives you raw packages with no validation, no version pinning, and no idea whether a skill helps or hurts. SkillzBench is an independent testing lab: it dedupes and verifies skills, benchmarks each in a sandbox with the same model with and without the skill, publishes the Lift in percentage points, and keeps money entirely out of scores. It also tracks upstream changes and re-scans, so you know when a skill you depend on shifts.

Agent Security Scanning for Content Creators