Skip to content
RosmarOps
  • Blog
  • About
  • Observatory
  • Sandbox
  • RosmarOps, the product
  • ES
  • CA

Blog

Can one AI agent change another agent’s configuration?

November 4, 2026October 29, 2026 by Albert Garcia Pujadas

Yes, if it is allowed to write where the other agent keeps its settings or instructions.

Categories CLAUDE.md, AGENTS.md and rules Tags AGENTS.md, Claude Code, Cursor

Which keys can your AI coding agent read?

November 4, 2026October 29, 2026 by Albert Garcia Pujadas

It depends on where the agent keeps its settings and on the mode it starts in.

Categories API keys and .env files Tags Claude Code, credentials, Cursor

Claude Code’s auto mode: what it changes about what your agent can read

October 29, 2026 by Albert Garcia Pujadas
An automatic door half open, an agent walking through toward folders, and a hand near a switch.

Auto mode is where recent versions of Claude Code start. What it approves on its own, what still asks, and the settings that narrow it, from the official docs.

Categories Approval modes Tags auto mode, Claude Code, permissions

RosmarOps 0.4.0: where your MCP servers send data, and when that changes

November 4, 2026October 28, 2026 by Albert Garcia Pujadas

RosmarOps 0.4.0 records where each MCP server sends data and warns when a server starts pointing somewhere else. Still no networking code.

Categories RosmarOps news Tags credentials, MCP, release

Before you tell your agent “install X from this repo”, look at what it leaves on your computer

November 4, 2026October 28, 2026 by Albert Garcia Pujadas

More tools now install by telling your agent to do it. What they leave behind is instructions and code your agents will follow. An example, file by file.

Categories MCP servers Tags AGENTS.md, Claude Code, skills

RosmarOps 0.3.0: Codex, Antigravity, Hermes Agent and Claude Desktop on the map

October 28, 2026October 28, 2026 by Albert Garcia Pujadas
RosmarOps logo

RosmarOps 0.3.0 detects four more AI agents, shows which of your agents can read the keys they keep, and is harder to fool. Still no networking code.

Categories RosmarOps news Tags Claude Desktop, Codex, credentials, release

RosmarOps is out: see what your AI coding agents can reach

November 4, 2026October 27, 2026 by Albert Garcia Pujadas

RosmarOps is out for Windows: a local, read-only map of what Claude Code and Cursor can reach on your machine. No networking code, no telemetry.

Categories RosmarOps news Tags Claude Code, Cursor, release

Where my 18/100 comes from: how the Agent Posture Score works

October 28, 2026October 27, 2026 by Albert Garcia Pujadas
A bar split into red, yellow and green zones, with weighted blocks pulling a marker into the red zone.

How RosmarOps turns what each AI agent can reach into a score out of 100: bands tied to the traffic light, public weights, and a versioned formula.

Categories Agent Posture Score Tags Agent Posture Score, Claude Code, Cursor, methodology

I audited my own AI agents. My machine scored 18/100.

November 4, 2026October 25, 2026 by Albert Garcia Pujadas
Illustration of an AI agent on a laptop reaching keys, files and other agents’ settings, next to a low score on a gauge.

I ran RosmarOps on my own machine. Claude Code scored 18/100: what it can read, what it can send out, and the one setting to change first.

Categories Agent Posture Score Tags Agent Posture Score, auto mode, Claude Code, credentials, permissions
Newer posts
← Previous Page1 Page2

Recent posts

  • The AI agent map: what each one can do without asking you
  • Your AI agent’s sandbox: it almost always has one, it’s almost never on
  • What’s new in RosmarOps: nine agents with a score, your projects across your whole folder and fewer interruptions
  • RosmarOps is now on the Microsoft Store
  • The real work starts when the agent says “I’m done”

© 2026 RosmarOps

© 2026 RosmarOps • Built with GeneratePress

Privacy · Legal notice · Terms · Cookie settings