Install
Please confirm you are human
This browser or connection looks automated. Press and continuously hold the control for 3 seconds to enable Google-hosted web results and, when separately allowed, AI-assisted answers.
A successful check enables 100 search requests. Interactive access does not authorize scraping, systematic collection, or reuse of search output.
News
The RubyGems agent attack is a coding-agent benchmark nobody writes
4+ hour, 18+ min ago (345+ words) This week the reports landed: an OpenAI agent swarm had been quietly uploading packages to RubyGems since early May. Over 2,000 packages on May 11-12 alone, naming themselves with "oai" prefixes, eventually forcing the registry to disable new-user signups for four days....
My privacy scan found 412 leaks. 374 of them were one line of my own tooling.
9+ hour, 31+ min ago (375+ words) Before publishing a repository I ran a scan for anything that should not go out. It found 412 hits in 16 files and I read that as sixteen files needing review. It was one defect, repeated. Every control in the build prints…...
The bug every AI coding tool ships, and how to prove it is gone
11+ hour, 2+ min ago (504+ words) The scanner is open source and the rules are readable before you trust a single finding: github.com/audit0/auditai-scanner. This post is about the one bug it was built for. Ask any AI coding tool for an invoicing app and…...
Do Coding Agents Really Need Unrestricted Shell Access?
16+ hour, 8+ min ago (1590+ words) Exploring a bounded repository capability layer with MCP, AST/LSP code intelligence, guarded writes, and Git context. Coding agents need access to a repository. They need to read files, search for implementations, follow definitions, change source, and inspect the resulting…...
Fixing AI Agent Supply Chain Attack: Node.js Blueprint
20+ hour, 9+ min ago (549+ words) This article was originally published on BuildZn. Key vectors for an ai agent supply chain attack: This is why ai agent attack prevention needs to be baked in from day one. You can't just trust the agent's "reasoning." My approach…...
Grad-CAM Told Me My Model Was Looking at the Right Thing. It Was Lying.
16+ hour, 48+ min ago (640+ words) Why It was lying, but only about one specific kind of shortcut. This is a full, reproducible teardown of when Grad-CAM catches …...
I built an epistemic gate to stop LLM data poisoning during fine-tuning. Tested across 5 architectures, orchestrated on a 2006 Toshiba laptop for $0.
21+ hour, 22+ min ago (450+ words) Hi everyone, For a long time, the AI industry has pushed the narrative that advanced safety research, data poisoning auditing, and model alignment require massive clusters and millions of dollars. I wanted to test if that's true. Over the last…...
The Hugging Face Incident: The AI That Hacked a Company to Steal Its Own Exam Answers
18+ hour, 18+ min ago (317+ words) In July 2026, an OpenAI model was given a cybersecurity benchmark to solve. Instead of solving it, it broke out …...
Benchmarking My MCP Security Scanner: 14 of 14 Findings, Zero False Positives
22+ hour, 48+ min ago (142+ words) A scanner without numbers is a demo. A scanner with made-up numbers is worse. So I hand-labeled a small benchmark suite and ran the scanner against it, 1000 times per fixture. The setup is deliberately boring: Aggregate: 14 of 14 planted findings detected,…...
Postgres MCP Pro Restricted-Mode Bypass Exposes the Gap in AI Database Security
1+ week, 1+ day ago (519+ words) Restricted mode in Postgres MCP Pro was designed as the safe configuration for exposing a PostgreSQL database to an AI agent. It limits operations to read-only transactions and validates incoming SQL against an allowlist. CVE-2026-85620, disclosed this week with a…...