ansible-safe-mcp
Provides tools to create, lint, plan, and apply Ansible playbooks against managed hosts, with guarded commit-confirmed execution and rollback.
Supports guarded Ansible changes on Linux hosts using on-host systemd timers and snapshots to automatically revert unconfirmed changes.
Supports guarded Ansible changes on macOS hosts using LaunchDaemon deadlines, snapshots, and rollback of files, defaults, and pf rules.
Supports MikroTik RouterOS changes through Safe Mode, holding the session during the change and rolling back if confirmation fails.
Supports guarded OpenWrt changes with an on-device timer that restores configuration, firewall, passwd, and related files if changes are not confirmed.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
Ansible Safe MCP
Run Ansible from Claude. Every change undoes itself unless it's confirmed.
Claude writes the playbook, shows you the diff, and runs it under a timer that lives on the host. To install it, give your agent this link:
https://github.com/mureev/ansible-safe-mcpWhy it's safe
The fear that stops people from letting an agent near their servers is specific: it changes something, loses the connection, and nobody can get back in. Ansible's own safety net can't catch that, because the rescue runs on the machine that just lost the connection. Network gear solved this long ago with commit confirmed: apply, start a timer on the device, and roll back unless someone reconnects and says "keep it". This brings that to every host Ansible can reach (Linux: a systemd timer; Windows: a SYSTEM task; macOS: a LaunchDaemon; RouterOS: Safe Mode; OpenWrt: a timer in RAM), and puts four more gates in front of it:
What a revert can't undo: deleted data, calls to outside services, and package downgrades. The plan says when a playbook touches those.
Related MCP server: openwrt-mcp
Is it for you?
You run a few servers, routers or PCs over SSH and want Claude to do the chores: patch, fix, look into why the disk filled up. It works on Linux, Windows, macOS, MikroTik RouterOS and OpenWrt, directly or through a jump host, from the Claude desktop app (chat or Cowork) or OpenAI Codex on a Mac. If you have Ansible Automation Platform, Red Hat's own MCP server fits better (the other options).
The timers also work without Claude: mureev.confirmed
is a plain Ansible collection on Galaxy.
For agents
Everything an agent needs, to install it and then to use it, is in AGENTS.md. Read it before installing.
License
MIT. Ansible® is a registered trademark of Red Hat, Inc. This project is not affiliated with or endorsed by Red Hat.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Let AI operate servers without SSH. Choose actions, approve risky changes, and audit every step.
Backups of servers, Docker and Zimbra to your own S3. Restore by chat; changes need 2FA approval.
Validates AI infra code on real VMs. Self-corrects until it works. No containers, no sandboxes.
Preventive human-approval write-gate for AI agents: writes commit only after a human approves.
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceEnables coding agents to read, query, and safely modify RouterOS 7 devices over the REST API, covering firewall, routing, DHCP, DNS, interfaces, containers, files, and diagnostics, with confirmation gates and a scheduled rollback guard in front of disruptive changes.40 PyPIMIT
- AlicenseNot gradedqualityBmaintenanceEnables an MCP client such as Claude Code to inspect and change a stock OpenWrt 25.12+ router over SSH, covering system status, logs, clients, firewall, packages, services, and WireGuard peers. Configuration edits are staged, checked, and armed with an automatic rollback that survives a reboot, all behind deny-by-default per-client grants, an audit log, and masked secrets.1MIT
- AlicenseNot gradedqualityFmaintenanceEnables AI clients to operate remote VPS hosts through a jump-host gateway with a fixed whitelist of tools, allowing them to read logs, inspect and edit configuration files, manage services, and deploy or roll back versions. Production writes require a per-client time-limited lease, and all actions are audited, verified, and automatically reverted on failure.MIT
- AlicenseBqualityBmaintenanceLets Claude manage Linux servers and Docker over SSH or locally, with previews and approval before every change.42MIT