prompt-injection
Safeguard articles tagged "prompt-injection" — guides, analysis, and best practices for software supply chain and application security.
135 articles
Prompt Injection Prevention: A Defense-in-Depth Guide
Prompt injection is the top risk on the OWASP list for LLM applications for a reason: there is no single patch. Preventing it means layering controls around a model that cannot reliably tell instructions from data.
AI Agent Tool-Use Security: Locking Down What Agents Can Do
The moment you give an LLM tools, it stops being a chatbot and becomes an actor in your systems. Tool-use security is about making sure a compromised agent hits a wall instead of a credential.
RAG Security Best Practices for 2026
Retrieval-augmented generation wired an untrusted-content pipeline straight into your model's context window. Here are the practices that keep a poisoned document or a leaked chunk from becoming an incident.
Clinejection: prompt injection turns AI coding bot into supply chain attack
A prompt-injected GitHub issue title hijacked Cline's AI triage bot, poisoned its build cache, and pushed a malicious npm release to 4,000 developers.
AI Agent Memory: Security Risks
Persistent memory makes AI agents more useful and more dangerous. A security engineer's walkthrough of how agent memory gets poisoned, exfiltrated, and weaponised, with concrete 2025 examples.
Black Hat Arsenal 2026 Preview: The Agentic AI and Supply-Chain Tools to Watch
Black Hat USA 2026 runs August 1–6 at Mandalay Bay, with Arsenal August 4–6. Here is an honest preview of the open-source tool categories worth your time — and how to tell signal from demo-day hype.
DEF CON 34 Preview: Agentic AI Security Takes Center Stage at Hacker Summer Camp
DEF CON 34 lands in Las Vegas August 6-9, 2026 under the theme 'Agency' — a deliberate nod to agentic AI. Here is what to watch, why it matters, and how to prepare before you board the plane.
OWASP Top 10 for LLM Applications explained
A breakdown of the 2025 OWASP Top 10 for LLM Applications—prompt injection, supply chain, excessive agency—with real examples and fixes.
Prompt injection attacks: direct vs indirect
Direct prompt injection comes from the chat box; indirect injection hides in the data your AI agent trusts. Here's how the two attack types differ and what stops each.
Agent hijacking: the real-world impact of prompt injection
From a zero-click Microsoft 365 Copilot breach to poisoned MCP servers, AI agent hijacking is now a real, documented software supply chain threat.
mcp-scan: detecting malicious MCP tool definitions
MCP lets AI agents call tools via plain-text descriptions the model trusts blindly. Here's how mcp-scan catches poisoning, rug-pulls, and shadowing.
Best LLM Security Tools in 2026: Guardrails, Red Teaming, and Runtime Defense Compared
An honest guide to the best LLM security tools in 2026 — from open-source guardrails and red-teaming scanners like NeMo Guardrails, garak, and LLM Guard to runtime APIs and full AI security platforms — with clear guidance on which job each one actually does.
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.