skills.rest
jfrog
jfrogOFFICIAL
Explore all skills by this author →
WebsiteGitHub

jfrog-mcp-management

Install, list, and remove MCP servers through the JFrog Agent Guard catalog.

1|Updated Aug 3, 2026
Search Tags:#MCP#configuration#governance#install#coding agents#jfrog#agent guard
One-click install
npx skills add https://github.com/jfrog/jetbrains-plugin --skill jfrog-mcp-management-jfrog
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: jfrog-mcp-management
Source: https://github.com/jfrog/jetbrains-plugin/tree/main/.junie/skills/jfrog-mcp-management
Command: npx skills add https://github.com/jfrog/jetbrains-plugin --skill jfrog-mcp-management-jfrog

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @jfrog/agent-guard, and includes scripts (resource) and references (resource) components.

What problem does it solve? Managing MCP servers across different AI coding agents is error-prone: each agent stores config in a different file, format, and scope, and organizations need governance over which MCPs developers may install. This Skill routes every MCP install, list, and removal through the JFrog Agent Guard so only catalog-approved MCPs are configured, with correct per-harness config entries. ## Core Features & Use Cases - Governed MCP installation: Discovers the approved MCP catalog via npx @jfrog/agent-guard --list-available, inspects a package's required inputs with --inspect, and writes a correct config entry for the detected agent (Claude Code, Cursor, Codex, VS Code, OpenCode, Devin, Kiro). - Agent Guard activation gate: Runs a mandatory Step 0 check to determine whether the organization enforces Agent Guard, and honors disabled or ambiguous states without bypassing governance. - Secrets-safe configuration: Never accepts secrets in chat; writes value references (${VAR}, ${input:id}, env_vars) per harness and handles OAuth login via --login with token caching in ~/.jfrog/jfrogmcp.conf.json. - Use Case: A developer asks their AI agent to "add the kubernetes MCP". The Skill checks Agent Guard status, shows the approved catalog, inspects the chosen MCP, writes the correct entry into the agent's config file, and guides enablement, restart, and verification that at least one tool is exposed. ## Quick Start Ask the agent to install an MCP server, for example: "Install the kubernetes-mcp-server MCP for this project using the JFrog Agent Guard."

Frequently Asked Questions about jfrog-mcp-management

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I install an MCP server with the JFrog Agent Guard?▼

Run the Step 0 activation check first, then list the approved catalog with `npx --yes --registry <REGISTRY_URL> @jfrog/agent-guard --list-available --project <KEY>`. Inspect the chosen package with `--inspect`, then write the config entry for your agent's harness and restart to verify tools appear.

How do I list which MCPs are approved or available for my project?▼

Use the Agent Guard `--list-available` command against your configured JFrog server and project key; that governed catalog is the authoritative answer. Do not rely on local IDE settings or allowlists, which do not reflect organizational approval.

Does the Agent Guard work with Claude Code, Cursor, and VS Code?▼

Yes, the Skill supports Claude Code, Cursor, Codex, VS Code, OpenCode, Devin, and Kiro, each with its own config file path, top-level key, and secret reference syntax. It detects the harness and reads exactly one matching harness reference file.

Why does my installed MCP show connected but expose zero tools?▼

A connected label only proves the Agent Guard proxy started, not that the upstream MCP responded. Zero tools usually means a required env var or header input was not exported in the launching shell, or its name case does not match the catalog input name.

Can I install an MCP manually instead of through the Agent Guard?▼

No. When the Step 0 check reports Agent Guard enabled, all installs must go through the approved catalog and Agent Guard proxy; manual install routes are a hard stop. If Agent Guard is disabled for the organization, normal install methods apply.

What are the requirements to run the JFrog Agent Guard commands?▼

You need Node.js 18 or newer on PATH, network access to the JFrog Agent Guard npm registry and platform, and either a configured jf CLI server or JFROG_URL plus JFROG_ACCESS_TOKEN environment variables. A valid JFrog project key is always mandatory.

Related Skills

Explore More Software Engineering Skills→
jfrogjfrog
5

JFrog Access

Manage JFrog Access tokens, users, groups, and permissions via REST endpoints.

Official
Intermediate
brikkoAIbrikkoAI
1

mcporter

Manage MCP servers, invoke tools, and generate TypeScript clients.

Official
Intermediate
jfrogjfrog
19

jfrog-package-safety-and-download

Check package safety and download packages from JFrog Platform via catalogs and remote repositories.

Official
Advanced
peerasak-upeerasak-u

add-mcp-server

Add and configure MCP servers in pi-config with connection verification.

Community
Intermediate