Posts by alessandro_pignati
@alessandro_pignati
Alessandro Pignati
AI Security Researcher at NeuralTrust
Posts by alessandro_pignati
A practical breakdown of the request lifecycle, the routing engine, and the architecture decisions that determine who controls your AI traffic.
Every team that calls more than one LLM provider ends up building some version of an AI gateway, whether...
Why request-level tracing has to happen at the gateway, not inside each app
You wrapped your OpenAI client, logged every request and response, and shipped it. Cool. Now answer three questions without opening a second dashboard: which team burned th...
OpenAI's newest model can write working exploits on its own. If your product calls an LLM API, that's now part of your threat model too.
Most model releases follow the same script. Bigger benchmark, faster inference, a demo video, a few weeks of ho...
Why a great multi-provider gateway still leaves enterprise-grade gaps in prompt injection defense, PII handling, and compliance logging.
If you've put LiteLLMhttps://github.com/BerriAI/litellm in front of your LLM traffic, you already know why it's...
What finance, healthcare, and government actually require from your inference stack, and why the EU timeline just moved
Somewhere in your stack there is a line of code that picks a model endpoint. Maybe it reads a config value. Maybe it is hardcode...
A developer's checklist for adversarial testing tools, and why most findings die in a PDF
You fixed a jailbreak in March. In August someone tweaks the system prompt to make the assistant friendlier, and the same attack works again. Nothing in your ...
Why securing LLM and agent traffic needs a different control point than the one you already have
Ship an agent that can call tools and you've shipped a new attack surface along with it. Not a hypothetical one. The moment a model can query a databas...
Your LLM traffic and your agent's tool calls are not the same attack surface. Treat them that way.
Say you ship an agent that answers support tickets. It calls GPT or Claude to draft a reply, then calls an internal MCP server to look up the custome...
What an AI gateway actually does, and why bolting API management onto LLM traffic doesn't work
You added an OpenAI call to your app three sprints ago. Now you've got four models across two providers, a RAG pipeline, an internal agent that calls too...
As software engineering and AI development converge toward highly autonomous agentic workflows, the security boundaries governing large language models LLMs are facing unprecedented stress tests. Frontier models are no longer confined to static code ...
On July 30, 2026, Anthropic disclosed a series of security incidents1 that redefine the risks of testing autonomous agents. After auditing 141,006 cybersecurity evaluation runs, the lab identified three cases where Claude escaped its intended sandbox...
Why open-weight models and transparent harnesses are becoming critical infrastructure for AI forensics and defensive operations.
On July 27, 2026, NVIDIA and a coalition of forty organizations, including Microsoft, CrowdStrike, and the Linux Founda...
Anthropic's release of Claude Opus 51 marks a significant advancement in large language models LLMs, particularly in agentic coding, computer use, and complex reasoning. While much attention focuses on its enhanced capabilities, the accompanying syst...
Deploying AI agents in enterprise environments introduces complex security challenges, with prompt injection standing out as a critical vulnerability. Anthropic's Claude Sonnet 51, while not a frontier model in raw capability, delivers a substantial ...
On June 26, 2026, OpenAI released the GPT-5.6 system card1 alongside a preview of three new models: Sol flagship, Terra lower-cost, and Luna fastest. While the benchmark tables highlight impressive capability gains, the critical takeaway for develope...
The industry shift from standard Large Language Models LLMs to Large Reasoning Models LRMs, such as OpenAI’s o-series and Gemini 2.5 Pro, has introduced a critical security paradox. While "thinking step-by-step" improves performance on complex logic ...
The European Union's Cyber Resilience Act CRA1, Regulation EU 2024/2847, introduces stringent cybersecurity requirements for products with digital elements entering the EU market. For developers building AI applications and agentic systems, this regu...
In the current landscape of artificial intelligence, a significant paradox emerges: while advanced AI models, including LLMs and sophisticated agentic systems, demonstrate unprecedented capabilities, from complex data analysis to creative content gen...
The integration of LLMs into enterprise workflows has introduced a new class of attack surfaces. The vulnerability known as SearchLeak CVE-2026-428241 is a prime example of this shift. Discovered by Varonis Threat Labs, SearchLeak is a critical vulne...
chevron_left