JFrog AI Agent Examples

A curated collection of AI agent skills and rules for automating workflows on the JFrog Platform. Each example is a self-contained set of skills that an AI coding assistant can follow to accomplish domain-specific tasks through natural-language conversation.

---

Disclaimer: This repository is a collection of community examples and is not a formal JFrog product. These skills and rules are provided as-is for interacting with the JFrog Platform and its features. They are not officially supported, may not cover all edge cases, and should be reviewed before use in production environments.

---

Overview

Example	Suggested persona	Description
compliance-and-policies	Platform / security admins, release managers	Create lifecycle policies that validate evidence before promotion
platform-features	Developers, DevOps, platform engineers	Multiple skills covering JFrog Platform APIs, CLI, security, and architecture patterns
onboarding-workflows	Platform admins, SRE, onboarding automation	Orchestrate multi-project onboarding with manifest-driven provisioning

Choosing an example: Use compliance-and-policies for evidence-based promotion gates; platform-features for API/CLI reference and patterns; onboarding-workflows for provisioning projects, repos, members, and CI from a manifest.

Repository structure

ai-agent-examples/
├── README.md                 # This file
├── .env.example             # Template for JFROG_URL, JFROG_ACCESS_TOKEN
├── global/                   # Shared rules applied across all examples
│   └── rules/
│       ├── interaction-questions.mdc
│       └── jfrog-platform.mdc
├── compliance-and-policies/  # Evidence-based lifecycle policies
│   ├── README.md
│   ├── rules/
│   └── skills/
├── platform-features/       # JFrog APIs, CLI, patterns
│   ├── README.md
│   ├── CONTRIBUTING.md
│   ├── rules/
│   └── skills/
└── onboarding-workflows/     # Manifest-driven project/repo onboarding
    ├── README.md
    ├── rules/
    ├── skills/
    ├── templates/
    └── scripts/

• global/rules/ — Copy these into your IDE’s rules directory so they apply whenever you use any example.
• <example>/skills/ and <example>/rules/ — Copy the ones for the example(s) you use.

Installation

Install skills directly from this repository using the skills CLI:

npx skills add git@github.com:jfrog/ai-agent-examples.git --skill <skill-name>

Replace <skill-name> with the skill folder name (e.g. jfrog-onemodel, jfrog-cli, etc.).

Note: Some skills depend on others (for example, jfrog-onemodel expects jfrog-cli for credentials) or on copied rules from global/ or an example’s rules/. Install those dependencies or copy the rules as needed.

Or copy the skills and rules for the example(s) you want into your AI environment. Run commands from the repository root (the directory containing global/, compliance-and-policies/, etc.).

Cursor

1. Copy global rules (recommended for all examples):

   mkdir -p .cursor/rules
   cp global/rules/*.mdc .cursor/rules/

2. Copy one or more examples:

   mkdir -p .cursor/skills
   # Example: compliance-and-policies
   cp -r compliance-and-policies/skills/* .cursor/skills/
   cp compliance-and-policies/rules/*.mdc .cursor/rules/ 2>/dev/null || true
   # Example: platform-features
   cp -r platform-features/skills/* .cursor/skills/
   cp platform-features/rules/*.mdc .cursor/rules/ 2>/dev/null || true
   # Example: onboarding-workflows
   cp -r onboarding-workflows/skills/* .cursor/skills/
   cp onboarding-workflows/rules/*.mdc .cursor/rules/ 2>/dev/null || true

VS Code (GitHub Copilot)

• Copy skill folders into .github/copilot/skills/ (or your Copilot skills path). Add rule content to .github/copilot-instructions.md or your project’s Copilot instructions so the agent follows JFrog conventions.

Claude Code

• Copy skill folders into .claude/skills/. Put global and example rule content into CLAUDE.md or a dedicated rules file your environment reads.

Windsurf

• Copy skills into .windsurf/skills/ and rules into .windsurf/rules/ (or the paths your Windsurf setup uses for skills and rules).

JetBrains AI Assistant

• Paste skill instructions into AI Assistant custom instructions or project guidelines (e.g. .junie/guidelines.md). Include the contents of the relevant rule files so the assistant follows JFrog API and safety conventions.

Other agents

• Skills use the open SKILL.md format. Copy the skills/ directories into your agent’s skills location. See agentskills.io for the broader ecosystem and tool support.

Prerequisites

• A JFrog Platform instance (entitlements depend on the example: e.g. AppTrust for compliance-and-policies, Platform Admin for onboarding-workflows).
• Credentials: set JFROG_URL and JFROG_ACCESS_TOKEN (or use a .env file). Copy .env.example to .env and fill in your values. Keep .env out of version control.

cp .env.example .env
# Edit .env with your JFrog URL and access token

• Tools used by the examples: curl, jq; for onboarding-workflows also yq, Git.

Using a skill

1. Install as above for your IDE.
2. Ask the agent in natural language, e.g.:
   • "Create a promotion policy that requires SLSA provenance evidence" (compliance-and-policies)
   • "Set up a CI integration with security scans for my npm project" (platform-features)
   • "Onboard all projects from my-manifest.yaml" (onboarding-workflows)

The agent will follow the skill’s workflow, use your credentials (from env or .env), and guide you step by step.

Contributing

We welcome new examples and improvements to existing ones.

1. New example: Add a directory at the repo root with:
   • README.md — overview, who it’s for, prerequisites, usage.
   • skills/ — at least one skill with SKILL.md (and optional *-reference.md, assets/).
   • rules/ (optional) — example-specific .mdc rules.
2. Existing example: Follow the same layout; keep shared conventions in global/rules/ and example-specific ones in that example’s rules/.
3. Skills: Include a clear desired state / success criteria and validation steps; use the standardized section order (Authentication, Prerequisites, Desired State, Workflow, Official Documentation, etc.) described in the plan.
4. Overlapping domains: If your skill touches Curation, AppTrust, Artifactory repos, or Access/projects, add a “Related skills” or “See also” reference to the other example(s) that cover the same end result.

License

This project is licensed under the Apache License 2.0 — see the NOTICE file for details.