News
CISA, G7 partners outline minimum AI SBOM elements
3+ day, 19+ hour ago (481+ words) CISA and G7 partners have released voluntary guidance outlining minimum elements for Software Bills of Materials tailored to artificial intelligence systems, giving public and private organizations a common structure for documenting AI components and dependencies. The document, "Software Bill of Materials…...
G7 Countries Release AI SBOM Guidance
4+ day, 41+ min ago (736+ words) The goal of the guidance, which outlines minimum elements, is to help organizations enhance transparency in AI systems and supply chains." Government agencies from Group of Seven (G7) countries this week published joint guidance to help organizations create a software bill…...
MCP governance for an AI coding agent without breaking the audit chain
4+ day, 3+ hour ago (934+ words) The Model Context Protocol gave AI agents a clean way to reach into systems. In a year it has become the default tool surface for serious agents. That is mostly good news. The mostly is the operative word. Without care,…...
Auditing LLM Agent Skill Integrity
4+ day, 10+ hour ago (248+ words) Startup Hub. ai Auditing LLM Agent Skill Integrity A new framework, Behavioral Integrity Verification (BIV), reveals 80% of LLM agent skills have implementation gaps, primarily due to oversight, and achieves 0. 946 F1 for malicious skill detection. The expansion of LLM agents into real-world…...
CISA, G7 Release AI Software Bill of Materials Guide
4+ day, 14+ hour ago (161+ words) Executive Gov The Cybersecurity and Infrastructure Security Agency and Group of Seven international partners have issued a new'joint guidance titled "Software Bill of Materials for AI " Minimum Elements." The new SBOM for AI guidance from CISA and the G7 reflects the…...
As agentic dev tools boom, workflow auditability becomes the constraint
5+ day, 23+ hour ago (353+ words) When AI coding agents open merge requests, audit trails often don't follow. Here's the compliance gap that's widening inside Dev Sec Ops teams....
Acceptance Cards Proposes Four-Diagnostic Standard for Safe Fine-Tuning
6+ day, 2+ hour ago (599+ words) According to an ar Xiv preprint (ar Xiv: 2605. 10575) submitted May 11, 2026, the paper introduces "Acceptance Cards": an evaluation protocol, documentation object, executable audit package, and claim-specific evidential standard for claims about safe fine-tuning defenses. The protocol requires passing four diagnostics-statistical reliability, fresh…...
Vertaa UX Integrations " Run UX Audits Anywhere
6+ day, 16+ hour ago (145+ words) Vertaa UX Run Vertaa UX wherever you work From the terminal to Figma to your AI agent's chat window. Every channel hits the same audit engine, the same WCAG 2. 2 checks, and the same scoring. - Browser Web App Browser-based dashboard for…...
In a Big Move to Linux Security, Debian Makes Reproducible Builds Mandatory
6+ day, 16+ hour ago (445+ words) Debian's release team has made reproducible builds a hard requirement for the Debian 14 "Forky" cycle. Since May 9, the project's migration software has blocked any package failing a reproducibility check from entering testing. If a package already in testing breaks reproducibility…...
append-only merkle trees for agent audit trails
1+ week, 1+ day ago (166+ words) nono published a blog post this week that's conceptually identical to what i built for mnemopay: an append-only merkle tree that records every action an agent makes, stored in a place the agent itself cannot reach. the key insight: if…...