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.
npx skills add Darwin-Agent/Car-bench-TRACE --skill windows-open-and-positiongit clone --depth 1 https://github.com/Darwin-Agent/Car-bench-TRACEWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/darwin-agent/car-bench-trace/windows-open-and-position)<a href="https://agentmods.dev/skills/darwin-agent/car-bench-trace/windows-open-and-position"><img src="https://agentmods.dev/badge/skills/darwin-agent/car-bench-trace/windows-open-and-position/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/darwin-agent/car-bench-trace/windows-open-and-position"><img src="https://agentmods.dev/badge/skills/darwin-agent/car-bench-trace/windows-open-and-position.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00155 | $0.03066 |
| Opus 5 | $0.00077 | $0.01533 |
| Sonnet 5 | $0.00031 | $0.00613 |
| Haiku 4.5 | $0.00015 | $0.00307 |
Grade A, and why
windows-open-and-position 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 8d 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.
How it starts
The opening of the file, as written. The whole thing — 79 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Open windows / sync window positions
Open or close windows to a position, or sync windows so they share a position. Requests range from a fully specified set (the percentage and which windows are given — just do it) to ones where the user leaves a real choice open (no percentage, or "sync the windows" without saying which to match to which). The same method handles both, plus the energy-efficiency check when opening past 25% with AC on, and the case where a control is missing.
When this applies
- "Open the windows / open all the windows for some air" — with or without a percentage.
- "Close all the windows" / "set the front windows to N%" — a fully specified open or close.
- "Sync the windows" / "get the windows to the same position" — with or without a stated which-to-which.
- "Close/set all windows to the same level as the passenger rear" — a match-to-a-readable-target sync.
- "Close the window that's open / fully open" — a single window identified by its current state; read positions to find which window matches the description and close only that one.
- Note: turning on AC, the front/rear window defrost, fan speed, or airflow direction are separate climate operations with their own skills — this skill only moves windows. Don't fan out into those unless the user asked for them.
- A common pairing is "close all the windows, then turn on the defrost / AC": the closing is squarely this skill — do it correctly here — but the climate side (defrost, fan, airflow, AC) and any auto-adjustments it triggers belong to that operation's skill, including reading and honouring the user's stored preference for those values rather than defaulting them. Don't pick a default airflow or fan value yourself to "help" the defrost or AC; that decision (and its preference lookup) is not part of moving windows.
- Do not add intermediate or temporary actuator changes that the user did not request, even as a "prerequisite" for the paired operation. Setting fan speed to some interim level so AC can be enabled — when the user neither asked for that level nor needs you to choose it — is an unrequested change that often gets immediately overridden by the user's next instruction; that intermediate actuation is a failure even when the final state ends up correct. Move only the windows; leave the paired climate tool to manage its own preconditions.
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.
- 8d ago First seen · 79 lines · 155 tokens per session scan A cec3ea3b1202
windows-open-and-position is a skill published in the GitHub repository Darwin-Agent/Car-bench-TRACE (8 stars, last pushed 5d ago), licensed MIT. It adds 155 tokens to every session and 3,066 once invoked, about $0.0008 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.
Other skills, from other repositories
gke-compute-classes
Configures, optimizes, and troubleshoots GKE ComputeClasses. Use when configuring Spot VMs with on-demand fallback, targeting specific accelerators (GPUs/TPUs) or machine families, restricting ComputeClass access, or debugging pending pods related to node pool auto-creation. Do not use for cluster-level Node Auto…
jetson-diagnostic
Read-only Jetson health snapshot for identity, memory, GPU, thermal, power, storage, services, and top processes.
doca-socket-relay
Use this skill when the operator is driving the DOCA Socket Relay to bridge a socket-oriented host application onto a BlueField DPU peer without rewriting it — picking the deployment shape (in-process, sidecar, or BlueField service container), configuring the host-side socket and the DPU-side forwarding endpoint…
offensive-z-wave
Z-Wave attack methodology — sniffing with Z-Force / EZ-Wave / RTL-SDR + ZniffMobile, S0 (legacy) network-key derivation flaw and key reuse, S2 (modern) ECDH commissioning analysis, replay/injection on unauthenticated nodes, default-key brute-force on test deployments, and home-automation hub pivots. Use when targeting…
hsb-flash
Flash the FPGA on an HSB board connected to an NVIDIA devkit. Supports HSB Lattice boards (FPGA versions 2407, 2412, 2507, 2510) and Leopard Imaging VB1940 "all-in-one" cameras (FPGA versions 2507, 2510). Uses release-specific YAML manifests and board-type-specific program commands. Lattice and VB1940 commands must…
jetson-validate-image
Use after jetson-flash-image to run static BSP checks, on-target smoke/regression tests on a flashed DUT, or both. Not for build or flash steps. Triggers: validate bsp, on-target validation.