AI
GitHub’s Open-Source AI Fuzzing Agent: How It Works
Continuous fuzzing does not scale on its own. Even projects enrolled in OSS-Fuzz for years hide critical bugs, and the reason is…
An MCP Server Workspace Escape, and the Bug Class It Invites
AWS disclosed a vulnerability in its own open-source MCP server on October 1, 2026, and the mechanism is worth more than the…
How AI Tool Calling Actually Works: Your Code Runs, the Model Only Asks
Tool calling in five steps: the model requests, your code executes, and the result goes back. Minimal Python pattern with guardrails.
Google Gemini Hacked 3 Companies During a Security Test: How to Secure AI Coding Agents
Google Gemini accessed three real companies during an authorized security test. Here is what happened and how to secure AI coding agents…
AI Coding Agents Are Changing Software Development: What Developers Need to Secure
Coding agents write branches, not main. Require reviews for sensitive files, scope tokens, enforce CI, and log agent traces.
How AI Agents Can Attack Software Supply Chains
Agents with publish tokens and no review can poison registries. What the reported RubyGems incident shows and how to lock down tokens,…
Ollama vs Cloud AI APIs: Cost, Privacy, and Performance
Local Ollama keeps data on your host with near-zero marginal cost. Cloud APIs scale with no ops but charge per token. How…
Why Frontier AI Development Is Becoming a Security Problem
By April 2026, two frontier models had solved multi-step network attack simulations end to end. The UK NCSC, NY DFS, Canada, and…
Run an Open-Weight LLM on Ubuntu With Ollama
Install Ollama on Ubuntu, run your first open-weight model with ollama run, manage local models, and call the API at localhost:11434. Tested…
How to Protect Git Repositories From AI Agent Abuse
Give an AI agent full repo access and it will use it. How to scope tokens, lock main, force human review, and…