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 DustinTrap/kvm-pilot --skill skillgit clone --depth 1 https://github.com/DustinTrap/kvm-pilotWrote 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/dustintrap/kvm-pilot/skill)<a href="https://agentmods.dev/skills/dustintrap/kvm-pilot/skill"><img src="https://agentmods.dev/badge/skills/dustintrap/kvm-pilot/skill/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/dustintrap/kvm-pilot/skill"><img src="https://agentmods.dev/badge/skills/dustintrap/kvm-pilot/skill.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 | $0.00231 | $0.02133 |
| Opus 5.5 | $0.00092 | $0.00853 |
| Sonnet 5.5 | $0.00046 | $0.00427 |
| Haiku 4.5 | $0.00023 | $0.00213 |
Grade A, and why
kvm-pilot 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.
How it starts
The opening of the file, as written. The whole thing — 146 lines — stays where its author put it; the contents beside it link to each section on GitHub.
kvm-pilot skill
⚠️ Verify against the matrix. The core read paths are live-verified on GL-RM1PE at beta maturity, but many device+capability combos are still unit-tested with mocks only (see the Hardware-Compatibility list for what actually has). Treat any combo the matrix does not show as live-verified as unverified, expect bugs there, and never point a destructive operation (power, reset, media, keystrokes) at a machine the user can't afford to have power-cycled unexpectedly. Surface each destructive step to the user before executing it.
This skill is a thin wrapper over the installable kvm-pilot package. The code
lives in the package, not here — install it and import it rather than copying
client logic into a script. Vision (screen classification) runs on Claude or
any local OpenAI-compatible VLM — see
references/setup.md.
How to use this skill — read the playbook at need-time
The detailed playbooks live in references/ next to this file, split so you
can read exactly the one the moment calls for instead of holding all of them
in mind. When one of these situations comes up, (re)read its file — even if
you read it hours ago. A fresh read at need-time beats a faded memory,
especially late in a long session:
| Situation | Read |
|---|---|
| Picking how to do an action (screen, HID, media, boot control, logs, SSH) — the full per-action interface matrix, multitasking rules, CLI command set | references/interfaces.md |
| Host dark / wedged / unreachable; a snapshot failed or can't be trusted; before ever asking the user to physically intervene | references/recovery.md |
| Installing the package, registering the MCP server, the full tool list + effect gates, approval cancel/denied handling, first-run orientation | references/setup.md |
| Any OS install through the KVM | references/linux-install.md |
| An installer or answer file asks for locale, keyboard, or timezone | references/target-context.md |
| Driving the Python library directly; worked examples | references/library.md |
What ships with it
6 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 3d ago Changed · +1 lines · -3 tokens per session b9d86ab0be7d
- 1mo ago First seen · 145 lines · 234 tokens per session scan A 965cb66387a9
kvm-pilot is a skill published in the GitHub repository DustinTrap/kvm-pilot (8 stars, last pushed 3d ago), licensed Apache-2.0. It adds 231 tokens to every session and 2,133 once invoked, about $0.0009 per session on Opus 5.5. 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
hardware-health-check
Out-of-band hardware health via Redfish BMC (read-only) — power state, component health, thermal and power readings, firmware inventory, SEL log triage. Use when determining whether a host is powered off versus unreachable, checking hardware faults, reviewing thermal or PSU state, or triaging BMC event logs.
iot-protocols
IoT protocol reference: selection matrix, MQTT QoS flows and LWT, topic design, CoAP observe and block-wise transfer, LwM2M objects and firmware update, BLE GATT and connection tuning, LoRaWAN airtime and classes, and Paho, Zephyr, and ESP-IDF client code. Use when designing or debugging the protocol layer of a…
rtos-patterns
FreeRTOS and Zephyr reference: priority inversion taxonomy, deferred interrupt processing, watchdog kick patterns, mutex vs. semaphore choice, stack sizing, FreeRTOSConfig.h settings, the stack overflow hook, and rate monotonic analysis. Use when designing, reviewing, or debugging RTOS firmware.
communication-buses
Bus reference for I2C, SPI, UART, CAN, and USB: addressing, pull-up sizing, bus recovery, SPI mode card, DMA cache rules, ring buffers, CAN bit timing and error states, USB descriptors, and STM32 HAL code for each bus. Use when writing or reviewing a bus driver or diagnosing a misbehaving bus.
cortex-m-patterns
Cortex-M code patterns for SysTick, NVIC priority grouping, FPU enable, MPU regions, AAPCS inline assembly, DWT cycle counting, vector table relocation, and WFI sleep entry. Use when writing core-level Cortex-M code or checking it against a known-good pattern.
bare-metal-patterns
Register-level C patterns for Cortex-M without a HAL: bit manipulation macros, GPIO and polled UART setup, SysTick timebase, interrupt-driven UART ring buffer, linker sections and map files, and weak default ISR handlers. Use when writing or reviewing HAL-free drivers and startup code.