RosmarOps began as a research notebook around one question: what can an AI system do that you don’t expect, and who decides what it reads. Today it is also a product that asks that question on your own computer.
Where it comes from
Since March 2025, the RosmarOps blog has published more than a hundred articles, in Spanish, along four lines:
- Adversarial SEO and AI SEO: how what search engines and assistants read, cite and recommend gets manipulated.
- Information warfare: how AI makes disinformation and influence operations cheaper.
- AI security: attacks on models, data poisoning and prompt injection.
- Agents and autonomy: what changes when AI stops answering and starts acting.
The thread that connects it
Across all of those topics the underlying question was the same: what gets into an AI system, and what can it do with it. From March 2026 the blog focused more and more on agents, because with AI coding agents that question reached the computer of the people who build: an agent reads files, runs commands and edits code with your permissions. To see how we got there (in Spanish):
- El agente que no sabe con quién habla (The agent that doesn’t know who it’s talking to)
- La capa de inferencia (The inference layer: when approving is no longer the same as deciding)
- MCP no es un plugin (MCP is not a plugin: agents’ new chain of trust)
What the product is today
RosmarOps reads the configuration of your AI coding agents (today Claude Code and Cursor) and draws an exposure map: what each agent can read and change, which keys it can reach, and whether one agent can change another agent’s configuration or instructions. Everything is computed on your computer. It doesn’t connect to the internet, doesn’t send anything to anyone, and doesn’t change anything on your machine; it only writes its own reports to its own folder. Today it evaluates Claude Code and Cursor on Windows, and it can’t be downloaded yet. More on the home page.
How to read this blog
The Spanish articles stay published as they were written: they are the archive of that research and reflect their moment. Product news is published here, on the RosmarOps blog, and in Spanish on the Spanish blog. The first post: I audited my own AI agents. My machine scored 18/100.