Posts from '2026'
AppSec Guide by Trail of Bits
Trail of Bits has a Testing Handbook with useful information about application security. Check it out:
Caveman mode for Claude Code consumes less tokens
A semantic constraint engine for Claude Code & Codex. Forces agentic communication into minimal-token lithic structures. Retain 100% technical accuracy while destroying up to 87% of output latency.
Learn the Claude Code Source
Well, you probably knows that the source code of Claude Code leaked today.
The most impressive is that someone created a 50 module learning tutorial about it.
Axios Compromised
Axios has been compromised. Versions 1.14.1 and 0.30.4 are malicious and inject a dependency on plain-crypto-js@4.2.1. Look for these versions and remove this dependency.
- Axios: https://security.snyk.io/vuln/SNYK-JS-AXIOS-15850650
- Plain-crypto-js: https://security.snyk.io/vuln/SNYK-JS-PLAINCRYPTOJS-15850652
Snyk Blog Post: https://snyk.io/pt-BR/blog/axios-npm-package-compromised-supply-chain-attack-delivers-cross-platform/
Hidden Layer Research
This site contains a lot of good research articles about AI Security.
AI Defend Framework
This link contains strategies to defend AI applications and integrations.
Run Claude Code using Ollama
This tutorial shows how to run claude code using a local Ollama model.
read more →
Maestro Threat Modeling Nanoclaw
This article details the result of a threat modeling using maestro framework in Nanoclaw by Nvidia.
https://kenhuangus.substack.com/p/maestro-threat-modeling-nemoclaw
Programming Fonts
This site shows a preview of each font for programming and terminal.
The AI Security Institute
This site contains a lot of researches about IA Governance and how AI Agents have been used in the real world.
Learn from incidents
This blog shares real cases about incidents and what we can learn from them
Trivy and LiteLLM was compromised
The Trivy GH Action was compromised and it affected a lot of other open source tools, creating an impactful attack chain.
Here is the Snyk Post: https://snyk.io/pt-BR/articles/poisoned-security-scanner-backdooring-litellm/
How you can be attacked by an MCP Server
A live session by José Augusto talking about MCP security
Marmite Blog Engine
I discovered this tool called Marmite to build blogs using markdown. It's awesome (I build this site using it hahaha).
Agent Security NotebookLM
I created a NotebookLM with all sources that I'm reviewing about security for AI Agents.