Tag: security
All the articles with the tag "security".
- llm-security 18 min read
Building an Agent That Can't Be Hijacked: The Design Rules
Eight frameworks tell you how to secure an agentic system and none of them are substitutes for each other. What they agree on, where they actively contradict each other, and the four properties I'm going to enforce in code in Part 2.
Read article - llm-security 26 min read
I Built the Defense, Then Attacked It: Numbers From a Secure Triage Agent
A capability-tracking interpreter for a SOC triage agent, a 20-case injection suite aimed at it, and 420 runs across three models. Exfiltration and disclosure go to zero and stay there regardless of model strength. Evasion tracks model capability, and one of my four policies turned out to be pure friction. Includes the LangGraph port, and why the framework's flagship agent pattern breaks the design.
Read article - llm-security 23 min read
Why a Classifier Won't Save You From Prompt Injection
Four documented ways to evade Prompt Guard-style detectors, and two underlying results — a cryptographic impossibility barrier and a measured security-fidelity tradeoff — that explain why this isn't a problem you fix with more training data.
Read article - ai 8 min read
Building RedTeam MCP: An AI-Powered Penetration Testing Assistant
A practical guide to building an MCP server that enables AI assistants to orchestrate offensive security tools for penetration testing, with proper safety guardrails
Read article - typescript 3 min read
TypeScript for Pythonistas: A Guide to Building Red Team Tools
A comprehensive TypeScript guide for Python developers. Learn the fundamentals by comparing Python and TypeScript code, focused on offensive security tool development and MCP servers.
Read article - adversarial-ml 14 min read
Adversarial Machine Learning: Attacks and Defenses
Deep dive into adversarial attacks against ML models: evasion, poisoning, and extraction. Exploring defenses, red teaming strategies, and the MITRE ATLAS framework for securing AI systems.
Read article