JFrog Skills
AI agent · Workflow / back-office agent · Operations & supply chain by JFrog
Procedural intelligence for your agentic software supply chain.
JFrog Skills provide procedural guidance and agent-driven workflows for the JFrog Platform, acting as a playbook for AI coding agents. They enable agents to perform complex DevOps tasks like artifact management and security reviews while ensuring compliance with organizational policies.
Agent specs
The full spec card (context, pricing per token, modalities and more) on AntSpace. Sign up free to see the full spec card.
- Languages
- English
Category: AI Agents
- Channels
- IDE, Terminal, In-app
- Deployment
- Cloud, Self-managed, On-premises, Hybrid
- Functionality
- Procedural guidance and task execution playbooks
- Primary Use Case
- Software supply chain operations and platform management
- Integration Method
- JFrog Agent Plugin or individual configuration
- Governance Tool
- Agent Guard and Skills Registry
- Procedural guidance for artifact search and build tracing
- Governed by Agent Guard to enforce organizational policies
- Bundled into the JFrog Agent Plugin for team-wide rollout
Overview
JFrog Skills are purpose-built AI capabilities that provide procedural guidance and agent-driven workflows across the JFrog Platform. Designed for developers and DevOps teams, these skills act as a "playbook" for AI coding agents, offering step-by-step instructions for complex software supply chain operations. They enable agents to perform tasks such as searching and downloading artifacts, tracing builds, and reviewing security or Curation signals without leaving the IDE or terminal. By providing procedural intelligence, they ensure that autonomous agents follow established organizational "golden paths" for software delivery. This integration reduces context switching and improves the accuracy of agent-driven DevOps tasks while maintaining governance through Agent Guard.
Agent Capabilities
- Platform Operations: Enables agents to complete administration tasks and repository management correctly using established playbooks.
- Security and Curation Review: Allows agents to review Curation signals and security status to ensure compliance with organizational policies.
- Supply Chain Intelligence: Gives coding agents the intelligence needed to follow an organization's existing golden path to software delivery.
Workflow Integration
- Plugin Integration: Bundled into the JFrog Agent Plugin for consistent rollout across development teams.
- Hybrid Intelligence Workflow: Works alongside the JFrog MCP Server to provide both procedural playbooks and live platform data.
- Multi-Agent Support: Supports popular coding agents including Claude Code, Cursor, OpenAI Codex, OpenCode, Kiro, and Devin AI.
Governance and Control
- Agent Guard Enforcement: Enforces organization policies natively inside the coding agent to control which skills can be invoked.
- Auditability and Traceability: Maintains a record of the MCP tools and skills agents invoke for visibility into agent activity.
- Registry Governance: Allows for the management of skill policies and waivers to govern the Skills Registry.
What it does
- Searching artifacts
- Downloading artifacts
- Tracing builds
- Reviewing security signals
- Reviewing Curation status
- Completing administration tasks
- Repository management
- Executing software supply chain operations
Use cases
- Developers use skills to prompt agents to handle repository management and project provisioning using natural language.
- Security teams use skills to allow agents to review Curation signals and security status for dependencies.
- DevOps engineers use skills to provide agents with step-by-step guidance for tracing builds and downloading artifacts.
- Organizations use skills to ensure AI agents follow established golden paths for software delivery across the team.
Specifications
- Channels
- IDE, Terminal, In-app
- Integrations
- Claude Code, Cursor, OpenAI Codex, OpenCode, Kiro, Devin AI, VS Code, IntelliJ, JetBrains IDEs, JFrog Artifactory, JFrog Xray
- Human control
- Agents follow procedural playbooks; Agent Guard enforces policies on which skills can be invoked; Skill waivers can be managed by administrators.
- Deployment
- Cloud, Self-managed, On-premises, Hybrid
- Setup
- Available via the JFrog Agent Plugin setup for organization-wide rollout.
- Functionality
- Procedural guidance and task execution playbooks
- Supported Agents
- Claude Code, Cursor, Codex, OpenCode, Kiro, Devin AI, VS Code, IntelliJ
- Primary Use Case
- Software supply chain operations and platform management
- Integration Method
- JFrog Agent Plugin or individual configuration
- Governance Tool
- Agent Guard and Skills Registry
- Data Access
- Artifact metadata, build info, and security signals via JFrog Platform
How it works
- Plugin Installation: Install the JFrog Agent Plugin in your preferred IDE or terminal environment to access bundled skills.
- Skill Invocation: The agent uses Skills as a playbook to understand how to perform complex JFrog Platform tasks correctly.
- Policy Enforcement: Agent Guard validates each skill usage against organizational rules before the agent executes the action.
- Task Execution: The agent executes the task, such as searching for a package or tracing a build, within the developer's workflow.
FAQ
What is the difference between JFrog Skills and MCP tools?
JFrog Skills provide procedural guidance (the playbook) for completing tasks, while MCP tools provide live platform data (the context) like package versions and vulnerability data.
Where are JFrog Skills managed?
The Skills Registry allows you to govern, manage policies, and handle waivers for the skills used across your organization.
Added by AntSpace from public sources (docs.jfrog.com).