Using Safeguard from Claude, Codex and ChatGPT: The Model Context Protocol (MCP) Server

IntermediateFree5 modules · 14 lessons · ~2 min

About this course

A hands-on course for developers who want to operate Safeguard from the AI assistant they already use. You will learn how the Model Context Protocol works, how to connect Claude, ChatGPT, the Codex CLI, Cursor, Copilot and Gemini to the Safeguard MCP server at mcp.safeguard.sh, and how to read risk, trigger scans and fix pull requests and check gates through tool calls without being misled by paging, scope or recalled facts. The course closes with least-privilege identity design, quota and troubleshooting patterns, and the governance features that keep an assistant's actions auditable.

What you'll learn

MCP Fundamentals for Safeguard Users

What MCP Is and Why Safeguard Ships a Server
The Safeguard MCP Surface: Endpoint, Directory and Tool Catalog
Authentication, Tenancy and Scope

Connecting Your Assistant

Connecting Claude: Web Connector, Claude Desktop and Claude Code
Connecting ChatGPT and the OpenAI Codex CLI
Cursor, Copilot, Gemini and the Safeguard Desktop App

Working Patterns

Reading Risk: Projects, Vulnerabilities, Findings and SBOMs
Acting Through the Assistant: Scans, Fix PRs and Gates
Prompting Patterns, Skills and Avoiding Hallucinated Security Facts

Operating Safely

01MCP Fundamentals for Safeguard Users3 lessons

What MCP Is and Why Safeguard Ships a Server
The Safeguard MCP Surface: Endpoint, Directory and Tool Catalog
Authentication, Tenancy and Scope

02Connecting Your Assistant3 lessons

Common questions

How much does Safeguard Academy cost?

Nothing. Every course, exam, and certificate on the catalog is free — including retakes. All you need is a free Safeguard account.

What are the prerequisites?

None. The flagship course, Safeguard Certified Practitioner, is a beginner-level course — basic familiarity with how software is built helps, but every exam question is answerable from the lessons themselves.

Least Privilege for Assistants: Identities, Tool Sets and Connector Permissions
Quotas, Rate Limits, Errors and Troubleshooting
Governing the Connection: MCP Guardrail, Agent Identity and Zero AI

Putting It Together

End-to-End Walkthrough: From a CVE Alert to a Merged Fix Through Claude
Team Rollout, Anti-Patterns and a Readiness Checklist

Connecting Claude: Web Connector, Claude Desktop and Claude Code

Connecting ChatGPT and the OpenAI Codex CLI

Cursor, Copilot, Gemini and the Safeguard Desktop App

03Working Patterns3 lessons

Reading Risk: Projects, Vulnerabilities, Findings and SBOMs
Acting Through the Assistant: Scans, Fix PRs and Gates
Prompting Patterns, Skills and Avoiding Hallucinated Security Facts

04Operating Safely3 lessons

Least Privilege for Assistants: Identities, Tool Sets and Connector Permissions
Quotas, Rate Limits, Errors and Troubleshooting
Governing the Connection: MCP Guardrail, Agent Identity and Zero AI

05Putting It Together2 lessons

End-to-End Walkthrough: From a CVE Alert to a Merged Fix Through Claude
Team Rollout, Anti-Patterns and a Readiness Checklist2 min

What happens if I fail the exam?

You can retake it after a 24-hour cooldown, as many times as you need. Retakes are free, and each attempt draws a fresh random set of questions.

How long is a certificate valid?

The Safeguard Certified Practitioner credential is valid for 24 months from issue. The expiry date is printed on the certificate and shown live on its public verification page. Renewing means passing the current exam again.