Agenticskills.io vs SkillzBench: Which Agent Skills Platform Should You Trust?
Agenticskills.io is a live, free directory of 193+ curated agent skills and 196 MCP servers across 9 platforms, backed by an 11-guide learning library — but it ranks skills with an editorial 'S-rank' rather than measured performance. SkillzBench, currently in pre-launch, is building the missing layer: sandbox benchmarking that tests every skill with and without the skill to publish a measured Lift, plus commit-SHA version pinning, security scanning, and a free MCP connector — all free.
TL;DR
- Agenticskills.io is live today with 193+ curated skills, 196 MCP servers, and 9 platforms tracked (as of Oct 13, 2026).
- SkillzBench is pre-launch — its waitlist/newsletter is live but the platform itself is not yet public.
- Agenticskills.io's S-rank is explicitly 'editorial judgment, not a measurement'; it does not publish with/without-skill benchmark results.
- SkillzBench plans to benchmark every skill in a sandbox with the same model and publish the Lift in percentage points, shown separately from ratings.
- SkillzBench plans to pin every skill version to a commit SHA and run license checks, security scans, and quality judging — relevant given published research finding ~26% of skills carry at least one vulnerability.
- Both are free: Agenticskills.io's directory and newsletter are free to use; SkillzBench plans to make its directory, MCP connector, benchmarks, and alerts all free.
- Agenticskills.io publishes an 11-guide learning library (8 topics, 41K+ words) covering Claude Skills, MCP servers, and the agentic ecosystem.
Head to head
| Dimension | SkillzBench | Agenticskills.io |
|---|---|---|
| Availability | Pre-launch; waitlist/newsletter live, product not yet public | Live now — 193+ skills, 196 MCP servers, 9 platforms indexed (Oct 13, 2026) 🏆 |
| Pricing | Everything free at launch: directory, MCP connector, benchmarks, alerts | Free to browse; newsletter 'free forever' |
| Skill verification | Planned commit-SHA version pinning, validation, license checks, security scans, quality judging 🏆 | Human curation and hand-audited guides; S-rank is editorial judgment, not a measurement |
| Benchmarking | Planned sandbox testing with the same model, with and without each skill, publishing Lift in percentage points 🏆 | None published — rankings are editorial |
| Security scanning | Planned per-skill security scans with re-scanning on upstream changes 🏆 | Security content and audit frameworks (e.g., analysis of 23,000+ public MCP servers), but no per-skill scanning on listings |
| Platform coverage | Skills for Claude Code, Codex, Cursor, Gemini CLI and chat-app agents; free MCP connects any agent | 9 platforms tracked, including Claude Code, Codex, Cursor, Gemini CLI |
| Learning resources | Blog in development; broad audience of business owners, developers, and agent builders | 11 guides, 8 topics, 41K+ words on skills, MCP servers, and workflows 🏆 |
| Staying current | Planned upstream change tracking, re-scanning, re-benchmarking, and user notifications | Live index updated with new skills and guides |
The contenders
AgenticSkills.io is a live directory and learning hub for AI agent skills. As of October 13, 2026 its index lists 193+ curated agent skills, 196 MCP servers, and 9 tracked platforms, covering agents that read SKILL.md files including Claude Code, OpenAI Codex, Cursor, Gemini CLI, and chat-app agents. Skills are organized into 16 categories and ranked with an editorial 'S-rank'; the site states listing is never paid for and that its rankings are editorial judgment rather than measurements. The site also hand-audits ecosystem content — for example, a 2026 security analysis of 8 independent audits covering 23,000+ public MCP servers found 73% fail basic security checks. Its learning library offers 11 in-depth guides across 8 topics totaling 41K+ words, on subjects from the npx skills CLI to the Agent Plugins open standard backed by OpenAI, Microsoft, AWS, GitHub, Cursor, and Vercel.
pricing See pricing
- + Fully live today — 193+ skills, 196 MCP servers, and 9 platforms indexed as of October 2026.
- + Broad platform coverage for any agent that reads SKILL.md, not just one vendor's tooling.
- + Substantial free learning content: 11 in-depth guides spanning the skills CLI, MCP servers, the Agent Plugins standard, and workflow stacks.
- + Transparent about methodology — the site explicitly labels its rankings as editorial judgment and states listings are never paid for.
- + Strong security awareness in its content, including a 12-step framework for auditing MCP servers before installing them.
- − S-rank is editorial judgment, not a measurement — the site does not publish with/without-skill performance benchmarks.
- − No commit-SHA version pinning or automated re-scanning of listed skills is mentioned on the directory.
- − No measured Lift or sandbox testing of skills is offered.
- − Directory listings for third-party tools show pricing only as a 'model, not a rate,' so cost comparisons require manual follow-up.
SkillzBench is AIQ Labs' free, independent directory of AI agent skills plus a free MCP tool that connects any agent to the skills you choose. It is currently pre-launch and actively being built — the waitlist/newsletter is live, the product itself is not yet public. SkillzBench is being built around five pillars: Discover (continuously crawl and dedupe skills from GitHub), Verify (pin every version to a commit SHA, validate, license-check, security-scan, and quality-judge), Benchmark (test every skill in a sandbox with the same model, with and without the skill, and publish the Lift in percentage points), Equip (one free MCP server plus per-agent loadouts where the agent only sees the skills and tools toggled on), and Stay current (track upstream changes, re-scan, re-benchmark, and notify). A trust score from 0 to 100 will combine security, quality, license, maintenance, and community signals, and Lift will always be shown separately from ratings — money never touches scores. The platform is being built in response to published research finding roughly 26% of skills carry at least one vulnerability.
pricing 0
- + Everything planned to be free — the directory, MCP connector, benchmarks, and alerts.
- + Measured benchmarking: every skill tested in a sandbox with the same model, with and without the skill, with Lift published in percentage points.
- + Lift always shown separately from ratings, and money never touches scores — revenue comes from advertising around blog articles only.
- + Commit-SHA version pinning plus license checks and security scanning directly address published research finding ~26% of skills carry at least one vulnerability.
- + Per-agent loadouts mean an agent only sees the skills and tools explicitly chosen for it, reducing exposure.
- + Continuous upstream tracking with re-scanning and re-benchmarking keeps skill assessments current.
- − Pre-launch: the product is not yet public, so none of its benchmarks, scans, or the MCP connector can be used today.
- − No published benchmark results, live skill counts, or live user counts exist yet.
- − As a newer entrant, it lacks an established track record compared to live directories.
Verdict
Winner: Agenticskills.io
Agenticskills.io is the usable choice today with a live index of 193+ skills and 196 MCP servers; SkillzBench is pre-launch and cannot yet be used, though its planned benchmark-and-verify model addresses what Agenticskills.io's editorial ranking does not measure.
| Option | Overall | Breakdown |
|---|---|---|
| Agenticskills.io | 7.5 | : 9: 8: 6: 4: 8 |
| SkillzBench | 6 | : 2: 10: 9: 9: 4 |
What you can actually use today
Editorial ranking vs. measured Lift
Cost and business model
Which should you pick?
- :
- :
- :
- :
- :
FAQ
Is Agenticskills.io free to use?
Yes — the directory of 193+ skills and 196 MCP servers, the learning library, and the newsletter are all free, with the newsletter explicitly described as 'free forever, unsubscribe anytime.' The site states listings are never paid for and that its S-rank is editorial judgment rather than a measurement.
Is SkillzBench available yet?
No. SkillzBench is pre-launch and actively being built — the waitlist/newsletter is live, but the directory, MCP connector, benchmarks, and alerts are not yet public. No live skill counts, user counts, or published benchmark results exist yet, so interested businesses should join the waitlist.
How do the two platforms rank skills differently?
Agenticskills.io uses an editorial S-rank — human curation it explicitly labels as judgment, not measurement. SkillzBench plans to benchmark each skill in a sandbox with the same model, with and without the skill, and publish the Lift in percentage points, alongside a 0–100 trust score from security, quality, license, maintenance, and community signals, with Lift always shown separately from ratings.
Why does skill security matter for small businesses?
Published research found roughly 26% of agent skills carry at least one vulnerability, and Agenticskills.io's own synthesis of 8 audits covering 23,000+ public MCP servers found 73% fail basic security checks. A small business installing an unvetted skill into a coding agent that touches real systems takes on that risk, which is why version pinning and security scanning matter.
Which platform should a small business choose?
Use Agenticskills.io today if you need a live directory and tutorials — it is the only one currently available. Join SkillzBench's waitlist if you want measured Lift benchmarks, security scans, and license checks before trusting skills; the two are complementary rather than mutually exclusive, and both are free.