Unified RouterOS search — start here for any question. One call runs an input classifier (command-path, version, device, topic, property) and FTS in parallel, returning pages plus classifier-informed side queries in a single response. Consolidates what used to require 3–5 separate tool calls. Res...
AI agents call routeros_search to retrieve information from Rosetta without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This is a pure read/query tool for searching MikroTik RouterOS documentation. It retrieves and classifies information without side effects. The parallel execution of an input classifier and full-text search, while computationally involved, remains a read operation. No data is created, modified, deleted, or code is executed on external systems.
From the tool's definition Tool performs 'searchable access', 'FTS retrieval', 'search', and 'returns pages' with 'top FTS matches'. Description emphasizes information retrieval ('start here for any question') with no mutation, deletion, or execution capabilities.
Documented attack patterns abuse exactly the kind of access routeros_search gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Rosetta, and nothing reaches the server without passing your rules. This is the rule we recommend for routeros_search:
{
"version": "1",
"default": "deny",
"tools": {
"routeros_search": {}
}
}routeros_search is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Unified RouterOS search — start here for any question. One call runs an input classifier (command-path, version, device, topic, property) and FTS in parallel, returning pages plus classifier-informed side queries in a single response. Consolidates what used to require 3–5 separate tool calls. Response shape: - classified: { version, topics, command_path, command_path_confidence, device, property } — what the classifier detected from your input; command_path_confidence is high/medium/low - pages: top FTS matches (title, path, URL, excerpt, best_section) - related: callouts, properties, changelogs, videos, commands, devices, skills, glossary — empty sections are omitted. Cap scales with \. It is categorised as a Read tool in the Rosetta MCP Server, which means it retrieves data without modifying state.
Register the Rosetta MCP server in PolicyLayer and add a rule for routeros_search: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Rosetta. Nothing to install.
routeros_search is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the routeros_search rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for routeros_search. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
routeros_search is provided by the Rosetta MCP server (tikoci/rosetta). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Rosetta, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
14 Rosetta tools catalogued and risk-classified — across an index of 43,000+ MCP servers.