Routeros Command Tree logo

Routeros Command Tree

Organization
tikoci
routeros-command-tree

RouterOS command tree introspection via /console/inspect API. Use when: building tools that parse RouterOS commands, generating API schemas from RouterOS, working with /console/inspect, mapping CLI commands to REST verbs, traversing the RouterOS command hierarchy, or when the user mentions inspect, command tree, RAML, or OpenAPI generation for RouterOS.

Overview

Publishertikoci
Repositoryrouteros-skills
Skill namerouteros-command-tree
Stars
85
Forks
18
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by tikoci on GitHub. Read the source before you install it.

Installation

Install the Routeros Command Tree AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/tikoci/routeros-skills/tree/main/routeros-command-tree
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/tikoci/routeros-skills.git /tmp/routeros-skills
mkdir -p .claude/skills
cp -r /tmp/routeros-skills/routeros-command-tree .claude/skills/routeros-command-tree
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Routeros Command Tree in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Routeros Command Tree on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Routeros Command Tree is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

RouterOS Command Tree & /console/inspect

Overview

RouterOS organizes all configuration and commands in a hierarchical tree. Every path in the CLI (like /ip/address/add) corresponds to a node in this tree. The /console/inspect REST endpoint lets you programmatically explore the entire tree — this is how tools like restraml (RAML/OpenAPI schema generator) and rosetta (MCP command lookup) build their databases.

The Command Tree Structure

RouterOS's command hierarchy has four node types:

Node TypeMeaningExample
dirDirectory — contains child paths/ip, /system
pathPath — a navigable level (often has commands)/ip/address, /interface/bridge
cmdCommand — an executable actionadd, set, print, remove, get, export
argArgument — a parameter to a commandaddress=, interface=, disabled=

Tree Example

text
/ (root dir)
├── ip/ (dir)
│   ├── address/ (path)
│   │   ├── add (cmd)
│   │   │   ├── address (arg) — "IP address"
│   │   │   ├── interface (arg) — "Interface name"
│   │   │   └── disabled (arg) — "yes | no"
│   │   ├── set (cmd)
│   │   ├── remove (cmd)
│   │   ├── get (cmd)
│   │   ├── print (cmd)
│   │   └── export (cmd)
│   ├── route/ (path)
│   │   └── ...
│   └── dns/ (path)
│       ├── set (cmd)
│       ├── cache/ (path)
│       │   ├── print (cmd)
│       │   └── flush (cmd)
│       └── ...
├── interface/ (dir)
│   └── ...
├── system/ (dir)
│   └── ...
└── ...

/console/inspect API

Endpoint

text
POST /rest/console/inspect

Requires basic authentication. Available on all RouterOS 7.x versions.

Request Types

RequestPurposeReturns
childList children of a pathArray of {type: "child", name, "node-type"}
syntaxGet help text for a nodeArray of {type: "syntax", text}
highlightSyntax highlighting dataPer-byte token classes — see the routeros-syntax-inspection skill
completionTab-completion suggestionsCompletion candidates — see routeros-syntax-inspection for validity checking

Listing Children

typescript
// List children of /ip
const children = await fetch(`${base}/console/inspect`, {
  method: "POST",
  headers: { ...authHeaders, "Content-Type": "application/json" },
  body: JSON.stringify({
    request: "child",
    path: "ip",
  }),
}).then(r => r.json());

// Response:
// [
//   { "type": "child", "name": "address",    "node-type": "path" },
//   { "type": "child", "name": "arp",        "node-type": "path" },
//   { "type": "child", "name": "cloud",      "node-type": "path" },
//   { "type": "child", "name": "dhcp-client", "node-type": "path" },
//   { "type": "child", "name": "dns",        "node-type": "path" },
//   { "type": "child", "name": "route",      "node-type": "path" },
//   ...
// ]

Getting Syntax Help

typescript
// Get description for /ip/address/add → address argument
const syntax = await fetch(`${base}/console/inspect`, {
  method: "POST",
  headers: { ...authHeaders, "Content-Type": "application/json" },
  body: JSON.stringify({
    request: "syntax",
    path: "ip,address,add,address",   // comma-separated path
  }),
}).then(r => r.json());

// Response:
// [{ "type": "syntax", "text": "IP address" }]

Path Format

The path field uses comma-separated segments (not slashes):

  • Root: "" (empty string)
  • /ip: "ip"
  • /ip/address: "ip,address"
  • /ip/address/add: "ip,address,add"
  • /ip/address/add → address arg: "ip,address,add,address"

When using the JavaScript Array.toString() method, this comma-separated format is produced naturally from an array: ["ip", "address", "add"].toString() → "ip,address,add".

Tree Traversal Pattern

To walk the entire tree recursively:

typescript
async function walkTree(path = [], tree = {}) {
  const children = await fetchInspect("child", path.toString());

  for (const child of children) {
    if (child.type !== "child") continue;

    const childPath = [...path, child.name];
    tree[child.name] = { _type: child["node-type"] };

    // For args, fetch the syntax description — but NOT inside dangerous subtrees
    if (child["node-type"] === "arg") {
      if (DANGEROUS_PATHS.some(p => childPath.includes(p))) continue;

      const syntax = await fetchInspect("syntax", childPath.toString());
      if (syntax.length === 1 && syntax[0].text.length > 0) {
        tree[child.name].desc = syntax[0].text;
      }
    }

    // Recurse into this child (child enumeration is safe even in dangerous subtrees)
    await walkTree(childPath, tree[child.name]);
  }

  return tree;
}

Dangerous Paths — Conservative Skip Policy

Old RouterOS versions can deadlock the REST server when scripting-keyword paths are queried for syntax/completion via /console/inspect. Crawlers conservatively skip syntax lookups inside subtrees containing any of these names:

text
where, do, else, rule, command, on-error

These are RouterOS scripting constructs. Treat the six-path list as a conservative crawler skip policy, not a timeless crash invariant: live probing in tikoci/restraml (MikroTik support case SUP-127641) isolated the confirmed deadlock to bare do with request=syntax/completion on RouterOS ≤7.20.8 — the other five paths return [] instantly there — and found it fixed by 7.21.4. (The original six-path folklore came from sequential probing: once do hangs the server, every path tested after it appears to crash too.) Enumerating children (child request) is safe on all tested versions, as are nested paths like ["do","command"].

The conservative approach (used in the example above) skips the entire arg when any ancestor matches a dangerous path, and pairs every request with a short timeout on old or unknown versions. The actual rest2raml.js implementation matches this pattern.

typescript
const DANGEROUS_PATHS = ["where", "do", "else", "rule", "command", "on-error"];

CLI Command → REST Verb Mapping

RouterOS CLI commands map to HTTP verbs in the REST API:

CLI CommandHTTP VerbREST URL PatternNotes
get (print)GET/rest/ip/addressReturns array of all entries
get (single)GET/rest/ip/address/*1Single entry by ID
addPUT/rest/ip/addressCreates new entry (not POST!)
setPATCH/rest/ip/address/*1Updates existing entry
removeDELETE/rest/ip/address/*1Deletes entry by ID
printPOST/rest/ip/address/printAction-style (also works as GET)
Other commandsPOST/rest/path/commandAction — reboot, flush, etc.

Key insight: REST PUT = create, PATCH = update. This is the opposite of many REST API conventions where PUT is idempotent update and POST is create.

RAML/OpenAPI Schema Generation

When generating API schemas from the command tree:

  1. Walk the tree to collect all paths, commands, and arguments
  2. For each cmd node:
    • get → generates both GET /path (list) and GET /path/{id} (single)
    • add → generates PUT /path with arg-based request body
    • set → generates PATCH /path/{id} with arg-based request body
    • remove → generates DELETE /path/{id}
    • Other commands → POST /path/command
  3. For each arg under a command, generate request body properties or query parameters
  4. The desc field from syntax lookups becomes the description

The .proplist and .query Parameters

All POST-based command endpoints accept two special parameters:

  • .proplist — selects which properties to return (like SQL SELECT)
  • .query — filter expression array (like SQL WHERE)

These are RouterOS REST API conventions, not standard REST patterns.

Output Formats

The inspect tree can be converted to multiple schema formats:

inspect.json (Raw Output)

The raw tree as returned by recursive /console/inspect calls. Each node has:

json
{
  "address": {
    "_type": "path",
    "add": {
      "_type": "cmd",
      "address": { "_type": "arg", "desc": "IP address" },
      "interface": { "_type": "arg", "desc": "Interface name" }
    },
    "set": { "_type": "cmd", ... },
    "print": { "_type": "cmd", ... }
  }
}

RAML 1.0 (schema.raml)

Converted to RAML 1.0 resource/method notation:

yaml
/ip:
  /address:
    get:
      queryParameters: ...
      responses: ...
    put:
      body:
        application/json:
          properties:
            address: { type: any, description: "IP address" }
    /{id}:
      get: ...
      patch: ...
      delete: ...

OpenAPI 3.0 (openapi.json)

Standard OpenAPI 3.0 schema generated from the same inspect tree (7.21.1+).

The inspect.json Data Model

Each version's inspect.json is the canonical source of truth for that RouterOS version's command tree. It captures:

  • Every navigable path in the CLI hierarchy
  • Every executable command at each path level
  • Every argument (parameter) for each command
  • Syntax descriptions for arguments

Tools can parse inspect.json offline without needing a live router — set INSPECTFILE env var and the schema generator will use the cached file instead of querying a router.

Common Patterns for Working with the Tree

Finding Commands at a Path

typescript
// Given an inspect.json node for /ip/address
const node = inspectData.ip.address;

// Commands are children with _type === "cmd"
const commands = Object.entries(node)
  .filter(([key, val]) => val._type === "cmd")
  .map(([key]) => key);
// → ["add", "set", "remove", "get", "print", "export", ...]

Finding Arguments for a Command

typescript
// Arguments of /ip/address/add
const addCmd = inspectData.ip.address.add;
const args = Object.entries(addCmd)
  .filter(([key, val]) => val._type === "arg")
  .map(([key, val]) => ({ name: key, description: val.desc }));
// → [{name: "address", description: "IP address"}, ...]

Traversing Directories

typescript
// Directories and paths (navigable children)
const children = Object.entries(node)
  .filter(([key, val]) => val._type === "dir" || val._type === "path")
  .map(([key]) => key);

Performance Notes

  • Full tree traversal takes many minutes against a live router (thousands of HTTP requests, each a separate POST to /console/inspect). With KVM acceleration the CHR responds quickly, but the sheer number of sequential requests adds up.
  • Each /console/inspect call is a separate HTTP request — no batch API
  • Use INSPECTFILE for development/testing to avoid repeated live queries
  • The tree is version-specific — different RouterOS versions have different command sets
  • Extra packages (container, iot, zerotier, etc.) add additional command tree branches

Additional Resources

  • MikroTik's new manual site has a browsable CLI Reference at https://manual.mikrotik.com/docs/CLI%20Reference/. It exposes command menus, flags, argument names, and types by package/path (for example, /docs/CLI%20Reference/system/ip/address). Treat it as the human-readable companion to /console/inspect; when it disagrees with a live router, /console/inspect from that router wins.
  • For REST API details: see routeros-fundamentals skill → REST API patterns
  • For running a CHR to query: see the routeros-qemu-chr skill
  • For /app YAML format (a feature visible in the tree under 7.22+): see the routeros-app-yaml skill

Frequently asked questions

What does the Routeros Command Tree AI skill do?

RouterOS command tree introspection via /console/inspect API. Use when: building tools that parse RouterOS commands, generating API schemas from RouterOS, working with /console/inspect, mapping CLI commands to REST verbs, traversing the RouterOS command hierarchy, or when the user mentions inspect, command tree, RAML, or OpenAPI generation for RouterOS.

Why use Routeros Command Tree on TypingMind?

Because you install it once and use it with any model. Routeros Command Tree is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Routeros Command Tree in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/tikoci/routeros-skills/tree/main/routeros-command-tree. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Routeros Command Tree?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Routeros Command Tree?

As many as you like. As long as a model supports skills, you can use Routeros Command Tree with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Routeros Command Tree AI skill free?

Yes. It is published on GitHub by tikoci under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇