AI Security
In-depth guides and analysis on ai security from the Safeguard engineering team.
779 articles
Agentic AI Supply Chain Vulnerabilities
Agentic AI systems trust tools and models at runtime, not build time. Real 2024-2025 incidents show how MCP servers and AI packages become supply chain attack vectors.
Agentic Identity and Privilege Abuse
AI agents now inherit more privilege than they need — and incidents like the Microsoft 38TB SAS-token leak and ServiceNow's Now Assist flaw show what happens when that privilege gets abused.
Agent Tool Misuse and Exploitation
Attackers don't need to hack AI agents — they just redirect their own tools. Here's how tool misuse works, real 2025 incidents, and how to stop it.
Agent Goal Hijacking: Redirecting Autonomous AI Objectives
Attackers are hijacking autonomous AI agents by planting instructions in content they read—no exploit needed. Here's how it works, real 2025 incidents, and defenses.
LLM Misinformation: Security Risks of Hallucinated Outputs
LLM hallucinations aren't just AI trivia — they invent packages attackers squat on, fake CVEs, and false advisories that have already cost real companies real money.
LLM Unbounded Consumption: Resource Exhaustion Attacks
How attackers exploit token-based pricing and growing context windows to exhaust LLM compute and inflate cloud bills — and the concrete limits that stop them.
LLM Vector and Embedding Weaknesses
Embeddings aren't anonymized math — Vec2Text recovers 92% of text from vectors, and OWASP's LLM08:2025 now names inversion, poisoning, and exposed vector DBs as core AI risks.
LLM System Prompt Leakage
System prompts often hide business logic and secrets. Here's how attackers extract them, real 2023-2024 incidents, and how to stop leaks before they reach production.
Model Theft: Protecting Proprietary LLMs from Extraction ...
A $20 API attack can clone a production LLM's embeddings. Here's how model extraction works, real incidents from LLaMA to DeepSeek, and how to protect proprietary models.
LLM Supply Chain Vulnerabilities
Malicious model files, poisoned datasets, and compromised ML packages are the new software supply chain frontier. Here is how these LLM attacks actually work.
Training Data Poisoning Attacks on Machine Learning Models
A $60 domain purchase or 0.001% of training tokens can silently corrupt an ML model. Here's how training data poisoning attacks work and how to defend against them.
Sensitive Information Disclosure in LLM Applications
From Samsung's ChatGPT leak to RAG pipelines with no access controls, sensitive information disclosure is now a top LLM security risk. Here's how it happens and how to stop it.
Self-healing security runs on Safeguard.
Your first fix PR is minutes away.
No sales call required, even your agent can complete the purchase over MCP.