What problem does it solve? Setting up the JFrog plugin involves many prerequisites — Node.js, the JFrog CLI, server credentials, MCP configuration, project selection, and AI Catalog entitlement — and a failure in any one of them leaves the environment broken with no clear diagnosis. This Skill walks an ordered checklist of detectors, stops at the first failing check, and guides the user through the exact fix before continuing. ## Core Features & Use Cases - Ordered readiness checklist: Runs eight detector steps (Node.js ≥ 18, jf CLI install, server connection, credential validity, MCP plugin entry, project resolution, AI Catalog entitlement, Claude marketplace registration) and stops at the first non-passing result. - Guided remediation: Each failure branch maps to an exact fix — approval-gated Node or jf installation, web-login vs. access-token authentication, and interactive server/project pickers. - Cross-harness support: Detects the host harness (Claude Code, OpenCode, Kiro, Junie, VS Code, Devin) and adapts MCP config paths and prompt mechanisms accordingly. - Use Case: After installing the JFrog plugin for the first time, run the initialization walk to install the JFrog CLI, connect it to your JFrog Platform via browser login, pick a project, and confirm the MCP server and AI Catalog are reachable — all in one guided pass. ## Quick Start Ask the agent to run /jfrog-init to verify and complete your JFrog plugin setup.