A

AgentSkillsHub

A directory of 1213+ security-scanned MCP skills, workflows, and integrations, each with an A–F security rating provided before installation.

Last Updated 10/10/2026

About AgentSkillsHub

AgentSkillsHub is a security-reviewed directory of MCP servers and AI agent skills, featuring 1213+ MCP skills, workflows, and integrations, of which 120 are official skills. Every skill is rated from A to F before installation, making it easy for developers using Claude Code and similar AI coding assistants to assess risk and make adoption decisions. Each skill entry includes installation instructions, a security score, and compatibility notes, and the site also offers a comparison workbench and security review guides to help teams systematically evaluate risks before bringing in agent tools.

Best for

Developers using Claude Code and similar AI coding assistants
Engineering leads who need to evaluate and adopt agent tools for their teams

Best use cases

Before installing a new MCP server or agent skill, check its security rating and scan results first

Compare source code, update time, installation path, and evidence boundaries across similar skills before making a choice

Follow the official security checklist to assess whether a newly integrated MCP server meets launch standards

Compare before choosing

According to its official comparison page, AgentSkillsHub and alternatives such as Skills.sh can be examined side by side in the comparison workbench, including source code, update time, installation path, and evidence boundaries.

Features

Security Scanning & A–F Ratings

Scans each skill's source code for 45+ vulnerability patterns, including arbitrary code execution, credential leakage, and unrestricted file access, and provides an A–F security rating before installation to clearly show the risk level.

Security Review Process

Provides a security review guide explaining how the rating model aggregates risk density based on implementation behavior, permission requirements, and scope of operational impact; and follows a repeatable review process: static pattern checks, behavior simulation, permission scope verification, and launch-readiness review.

Comparison Workbench

In the comparison page, you can examine skills' source code, update time, installation path, and evidence boundaries side by side before deciding whether to install.

Skill Directory

Includes 1213+ MCP skills, workflows, and integrations covering common use cases such as GitHub integration, code review automation, deployment pipelines, Stripe payment processing, invoice generation, and expense tracking.

Pros

  • Every skill receives an A–F security rating before installation, so risk levels are clear at a glance
  • Scans skill source code for 45+ vulnerability patterns, covering common risks like arbitrary code execution and credential leakage
  • Offers a repeatable review process and a comparison workbench, making it easy for teams to evaluate and compare skills consistently

FAQ

AgentSkillsHub is a curated directory of MCP servers and AI agent skills for developers using Claude Code and similar AI coding assistants. Each skill entry includes installation instructions, a security score, and compatibility notes, helping teams evaluate and adopt agent tools.
Yes. The official security guidance explicitly states that the score is only a risk signal, not an absolute guarantee of safety; you still need to review the permission scope and deployment environment.
According to the official guidance, the rating label summarizes risk density rather than feature richness: the score takes into account implementation behavior, permission requirements, and the scope of operational impact. A skill with fewer features but strict boundaries may be safer than one with many features.

Tool Information

DeveloperAgentSkillsHub
Release Date10/10/2026
PricingFree