packet-tracer-mcp 0.9.0

Tell your AI "create a network with 3 routers, OSPF and DHCP" — it plans, validates, generates, and deploys the topology directly into Cisco Packet Tracer in real time.

61 MCP Tools

5 MCP Resources

74 Device Models

151 Modules

15 Cable Types

🌐 Website: https://www.mcpnetwork.top  •  📚 Documentation: https://mats2208.github.io/MCP-Packet-Tracer/

Showcase

_{3-router linear topology with OSPF, DHCP, and 6 PCs — planned and deployed via MCP tools}

_{Live deploy — from a natural-language prompt to a running topology in Packet Tracer}

What it does

A Model Context Protocol (MCP) server that gives any LLM (Claude, GitHub Copilot, Codex, …) full programmatic control over Cisco Packet Tracer.

👉 Full tool reference, device catalog, networking guides and architecture live in the documentation site.

Installation

1. Install the server

git clone https://github.com/Mats2208/MCP-Packet-Tracer
cd MCP-Packet-Tracer
pip install -e .

2. Connect your MCP client (Claude Code shown)

Linux · macOS · Git Bash · Windows cmd.exe:

claude mcp add --scope user --transport stdio packet-tracer -- python -m packet_tracer_mcp --stdio

Windows PowerShell — quote the -- separator, or PowerShell swallows it and Claude aborts with error: unknown option '-m':

claude mcp add --scope user --transport stdio packet-tracer "--" python -m packet_tracer_mcp --stdio

Verify with claude mcp list (look for packet-tracer … ✓ Connected).

3. Install the live-deploy extension — only if you want real-time deploy into a running Packet Tracer

Download V5.pts from Releases, then in Packet Tracer go to Extensions → Scripting → Configure PT Script Modules → Add… and select it. Full walkthrough in Live deploy below.

v0.6.0+ requires V5. The bridge now authenticates with a per-machine token that the V5 extension reads automatically; builds before V5 can't authenticate.

4. Install the Claude Code Skill — recommended; makes the AI use the MCP correctly instead of guessing

The repo ships a companion Agent Skill that teaches the model the exact tool catalog, the discover→plan→validate→deploy workflow, and the precise Script-Engine API (so it never invents method/model/port names). Install it globally from the repo root:

Linux · macOS · Git Bash:

mkdir -p ~/.claude/skills/packet-tracer && cp skill/SKILL.md ~/.claude/skills/packet-tracer/SKILL.md

Windows PowerShell:

New-Item -ItemType Directory -Force "$HOME\.claude\skills\packet-tracer" | Out-Null; Copy-Item skill\SKILL.md "$HOME\.claude\skills\packet-tracer\SKILL.md"

Then run /reload-skills in Claude Code (or restart it) and confirm with /skills. Details → Skill docs.

Requires Python 3.11+ (deps mcp[cli]>=1.13, pydantic>=2.11 install automatically). Full setup for every client → Installation docs.

Quick start

Just talk to your AI:

"Build a network with 2 routers, 2 switches, 4 PCs, DHCP and static routing."

The LLM calls pt_full_build, which plans → validates → generates → deploys. See the Quick Start guide.

Live deploy

Stream topologies into a running Packet Tracer in real time. Install this repo's own MCP Control Center extension once — the .pts from Releases — via Extensions → Scripting → Configure PT Script Modules → Add…, then open Extensions → MCP BUILDER. It auto-connects to the bridge — no snippet to paste.

_{Installing the MCP Control Center extension (V5) in Packet Tracer}

📖 Full steps → Live Deploy Setup.

Clients

Any MCP client drives this server — Claude Code, Cursor, Claude Desktop, VS Code with Copilot, Codex. Nothing in it is client-specific.

There is also one built on it: PacketSmith, a terminal app that runs these 61 tools with the network drawn beside the conversation — a fabric tree and a canvas plan derived from the pt_* results themselves, so a device the model says it created but did not never shows up.

You need this server either way — PacketSmith runs it underneath, and its setup installs it for you. If you already live in Claude Code, you are done here; the client is for when you want the topology in front of you instead of buried in a scrollback.

npm i -g packetsmith    # beta · MIT

Credits & Acknowledgements

Live deploy runs through our own Packet Tracer extension — the MCP Control Center (the .pts in Releases). Its Script-Engine helper layer was inspired by PTBuilder by Kim Knight (@kimmknight), who pioneered driving Packet Tracer's Script Engine from JavaScript — thanks for the groundwork. 🙏

PTBuilder and Packet Tracer MCP are separate, independent projects. You install our extension, not PTBuilder. Full Credits & Attribution.

Security

Driving Packet Tracer from outside means running a local HTTP bridge whose whole job is to hand JavaScript to PT's Script Engine — code that executes with PT's own privileges, including disk access. That makes the bridge a genuine attack surface, not an implementation detail, and it is hardened accordingly.

The attack this design exists to stop. Binding to 127.0.0.1 is not a security control. A POST with Content-Type: text/plain is a CORS simple request: any web page open in your browser can send it to a loopback port without a preflight and without needing to read the response. An unauthenticated bridge therefore lets any website you visit — while Packet Tracer happens to be open — queue arbitrary code inside it. Injection never needed to read anything back, so same-origin policy alone never closed this.

What actually closes it is a secret the attacking page cannot guess:

Regression coverage lives in tests/test_bridge_security.py and tests/test_injection_regressions.py; the full suite runs offline with python -m pytest — no Packet Tracer required.

v0.6.0+ requires the V5 extension. Versions before v0.6.0 shipped an unauthenticated bridge and are vulnerable to exactly the attack above. Upgrade — there is no safe configuration of the old bridge.

Deliberate, documented behaviour: pt_send_raw executes arbitrary JavaScript inside Packet Tracer by design — it is the escape hatch for exploring the IPC API. It is reachable only by an MCP client you have already authorised, over the authenticated bridge. That is a capability, not a vulnerability.

Found a vulnerability? Report it privately via GitHub Security Advisories, not a public issue. SECURITY.md documents the full threat model.

What's new

v0.8.0 — the agent can now show the network, not just describe it: canvas screenshots plus notes and drawings, for topologies that document themselves. v0.7.0 made the server read a live topology, not just build one: security auditing, per-port inspection, packet tracing with Packet Tracer's own per-layer decision log, NetFlow, and config backup. It also fixes pt_full_build(deploy=True), which used to report success while leaving the canvas empty. Full list in the Changelog.

Contributing

See CONTRIBUTING.md. Tests run offline with python -m pytest; no Packet Tracer needed.

License

Released under the MIT License — © 2026 Mateo (@Mats2208).