AI & Agent Security with Safeguard

IntermediateFree10 modules · 19 lessons

About this course

How Safeguard secures AI as a first-class object — model scanning, AI-BOM discovery, the AI Gateway firewall, MCP Guardrail — and how the platform's own purpose-built models (Griffin, Eagle, Lion/Lino) and agent-native design fit together.

What you'll learn

Safeguard's Own Models

Griffin, Eagle, and Lion/Lino
AI Modes and the Privacy Posture

Securing AI as an Asset

AI-BOM, AI-SPM, and Model Scanning
AI Gateway and MCP Guardrail

Agent-Native Operation

MCP Server and In-App Agentic Search
Autonomous Agent and Agentic Procurement

AI-BOM: Inventorying What You Actually Run

Discovering AI Assets Across Code, Cloud, and SaaS
Model Artifact Risk: Formats, Scanning, and Safe Adoption

Runtime Defence for LLM Applications

Prompt Injection: A Defender's Taxonomy
Data Egress, Secrets, and the DSPM Connection

Agent Governance

Agent Identity, Tool Access, and MCP Governance
AI Compliance: EU AI Act, ISO 42001, and NIST AI RMF

Threat Modelling LLM Applications

OWASP Top 10 for LLM Applications: Mapping Risks to Controls
Detecting Indirect Injection: Provenance Tagging, Canaries, and Response-Path Monitoring

Jailbreak Resistance and Output Safety

Jailbreak Resistance: Pattern Classes, Session-Level Scoring, and Measuring Robustness
Output Handling: Schema Enforcement, Argument Binding, and Grounding

Model Supply Chain Defence

Inside Model Artifact Scanning: Pickle, PyTorch, safetensors, GGUF, and ONNX
Poisoning Defences: Lineage, Evaluation Gates, and Retrieval Corpus Controls

Securing MCP and Agent Tooling

MCP Threat Model: Tool Poisoning, Rug Pulls, and Guardrail-Proxy Mitigations

01Safeguard's Own Models2 lessons

Griffin, Eagle, and Lion/Lino
AI Modes and the Privacy Posture

02Securing AI as an Asset2 lessons

AI-BOM, AI-SPM, and Model Scanning

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.

AI Gateway and MCP Guardrail

03Agent-Native Operation2 lessons

MCP Server and In-App Agentic Search
Autonomous Agent and Agentic Procurement

04AI-BOM: Inventorying What You Actually Run2 lessons

Discovering AI Assets Across Code, Cloud, and SaaS
Model Artifact Risk: Formats, Scanning, and Safe Adoption

05Runtime Defence for LLM Applications2 lessons

Prompt Injection: A Defender's Taxonomy
Data Egress, Secrets, and the DSPM Connection

06Agent Governance2 lessons

Agent Identity, Tool Access, and MCP Governance
AI Compliance: EU AI Act, ISO 42001, and NIST AI RMF

07Threat Modelling LLM Applications2 lessons

OWASP Top 10 for LLM Applications: Mapping Risks to Controls
Detecting Indirect Injection: Provenance Tagging, Canaries, and Response-Path Monitoring

08Jailbreak Resistance and Output Safety2 lessons

Jailbreak Resistance: Pattern Classes, Session-Level Scoring, and Measuring Robustness
Output Handling: Schema Enforcement, Argument Binding, and Grounding

09Model Supply Chain Defence2 lessons

Inside Model Artifact Scanning: Pickle, PyTorch, safetensors, GGUF, and ONNX
Poisoning Defences: Lineage, Evaluation Gates, and Retrieval Corpus Controls

10Securing MCP and Agent Tooling1 lessons

MCP Threat Model: Tool Poisoning, Rug Pulls, and Guardrail-Proxy Mitigations

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.