unity-driving

Instructions for controlling a running Unity development build through the Mono Soft Debugger, a protocol for inspecting and changing a game's live program state. It applies to development builds, not ordinary retail builds.

In plain words
What is it for?
Use it before driving a Unity game with debugging tools, especially for checking the live connection or planning several state changes.
Why use it?
It explains how to attach safely, understand the current debugger connection, and diagnose connection problems such as blocked multicast traffic.

Skill for Claude CodeCodex

Part of the unity-devtools plugin — 1 skill shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/citiesskylinesmodding/agents-plugins/unity-driving
Any agent
npx skills add CitiesSkylinesModding/agents-plugins --skill unity-driving
Clone the repo
git clone --depth 1 https://github.com/CitiesSkylinesModding/agents-plugins

Made for: Claude Code, Codex.

Or install unity-devtools, the plugin that ships this one along with the rest of its 1 skill.

Per session 46 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,188 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00046 $0.03188
Opus 5 $0.00023 $0.01594
Sonnet 5 $0.00009 $0.00638
Haiku 4.5 $0.00005 $0.00319

Measured 3d ago against content hash add015e2f40c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

unity-driving scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/unity-devtools/skills/unity-driving/SKILL.md · 113 lines

How it starts

The opening of the file, as written. The whole thing — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Driving a Unity game over SDB

The unity tools drive a running Unity Mono development build over the Mono Soft Debugger protocol (SDB); this skill is the procedure the tool schemas cannot carry. Everything holds for any dev-Mono Unity game; the type names in examples are placeholders (MyGame.…), so substitute the target's own. A retail build exposes no SDB port; only a development Mono build is drivable.

Session lifecycle

The first tool that needs the VM attaches on its own, the session persists, and a dropped connection reattaches on the next call, so just call the tool you need. The game is found by the PlayerConnection beacon it multicasts, which carries the debugger endpoint: there is nothing to configure, and every attach resolves that endpoint afresh, so a game restart costs nothing. status is the read-only orient step: what the beacon advertises, the endpoint an attach would use, and session state (attached, held suspensions). Everything it reports describes the present, so act on it as read. A game running while status reports no beacon means the beacon is not getting through: multicast blocked by a firewall or a VPN interface, or beaconFault/beaconListening reporting the listen itself impaired. Either way, call attach with the game's debugger port. That port governs that one attach, so give it again after any reattach for as long as the beacon still cannot see the game. Rule that out first when status also shows a held suspension: a suspended game stops broadcasting, so a window held past ten seconds makes its own beacon expire. Resume and ask again before diagnosing the network. A beacon whose debuggerEnabled is false is a game running without player-connection-debug=1: only a relaunch with that option makes it attachable. The debugger slot is exclusive: while attached, an IDE debugger (Rider/dnSpy/VS) cannot attach to the game, and vice versa; detach frees the slot, and the next unity call reattaches on its own. An attach failure while an IDE holds the slot looks like a connection refusal, not "slot taken". A Unity Editor is attachable on the same protocol, but no Editor has been confirmed on the beacon: try status first, and where it reports none, find the Editor yourself by reading its debugger port off the OS listener table for that process and passing the port to attach. Expect a domain reload — a script compile, entering or leaving play mode — to drop the connection.

Read the full file on GitHub · 113 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 113 lines · 46 tokens per session scan A add015e2f40c

Subscribe to this mod's changes

unity-driving is a skill published in the GitHub repository CitiesSkylinesModding/agents-plugins (5 stars, last pushed 3d ago), licensed MIT. It adds 46 tokens to every session and 3,188 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories