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.
git clone --depth 1 https://github.com/air-gapped/skillsWrote 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/plugins/air-gapped/skills/inference-host)<a href="https://agentmods.dev/plugins/air-gapped/skills/inference-host"><img src="https://agentmods.dev/badge/plugins/air-gapped/skills/inference-host/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/plugins/air-gapped/skills/inference-host"><img src="https://agentmods.dev/badge/plugins/air-gapped/skills/inference-host.svg" alt="Reviewed on agentmods" width="80" height="20"></a>Grade A, and why
inference-host 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 16d 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.
What it actually says
{
"name": "inference-host",
"source": "./",
"description": "Inference host bring-up + tuning suite — Linux/GPU bare-metal host work that sits beneath any inference framework (vLLM, sglang, TensorRT-LLM). Two phases. Bring-up (nvidia-datacenter-bringup): Ubuntu 24.04 LTS from clean OS to gpu-operator cuda-validator passing — B300/B200/H100/A100/L40S/L4 driver + fabricmanager + NVLSM + DOCA-OFED install order, NVIDIA CUDA repo + DOCA repo + air-gap mirror, MOK + DKMS sign-on-build under Secure Boot, Dell PowerEdge XE9780/XE9785 baseboard firmware via iDRAC Redfish DellOemChassis.ExtendedReset, gpu-operator pre-installed-driver-mode integration. Tuning (gpu-host-tuning): read-only snapshot of CPU power state, C-states, NUMA topology, PCIe link state, GPU settings, kernel boot params, sysctl, ulimits, IRQ affinity, container runtime; optional pinned-host↔GPU memcpy bench (torch + numactl); per-lever cheat-sheets to flip settings (governor, EPP, cpuidle, persistence, ECC, hugepages, intel_iommu, NCCL env, tuned-adm profiles, BIOS guidance for Dell XE / SMC / HPE). Surfaces config gaps that bottleneck LMCache CPU-tier throughput, KV offload, NCCL bandwidth, prefix-cache rebuild.",
"version": "0.20261003.25",
"author": {
"name": "Jörgen"
},
"license": "MIT",
"skills": [
"./.claude/skills/nvidia-datacenter-bringup",
"./.claude/skills/gpu-host-tuning"
],
"strict": false,
"category": "inference",
"tags": [
"gpu",
"host-tuning",
"nvidia",
"hgx",
"dgx",
"b300",
"blackwell",
"fabricmanager",
"nvlsm",
"doca-ofed",
"secure-boot",
"mok",
"dkms",
"ubuntu-24.04",
"air-gap",
"dell-poweredge",
"idrac",
"numa",
"pcie",
"bios",
"nccl",
"tuned-adm",
"performance",
"bare-metal"
]
}What it installs
The manifest is a name and a version. 2 skills travel with it, and installing the plugin installs all of them — 437 tokens a session between them. Each is measured on its own page, and each can be installed alone.
What ships with it
1 file beside marketplace.json#inference-host 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.
- 16d ago Changed 829d8699966a
- 28d ago Changed af2398a1bcea
- 1mo ago Changed cdd8fb0359cf
- 1mo ago Changed 76fa36d16fd2
- 1mo ago First seen · 42 lines scan A 64ded5913d0b
inference-host is a plugin published in the GitHub repository air-gapped/skills (5 stars, last pushed yesterday), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. 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 2026-09-03.
Other plugins, from other repositories
infra-plugin
Infrastructure management for Kubernetes/GKE, Istio service mesh, and authentication (Keycloak, OAuth2-proxy). Includes Helm charts, Terraform IaC, and Traefik reverse proxy.
deployment-validation
Pre-deployment checks, configuration validation, and deployment readiness assessment.
systems-architecture
Systems architecture and distributed systems skills including Designing Data-Intensive Applications (Martin Kleppmann), System Design Interview (Alex Xu), Clean Architecture (Robert C. Martin), Release It! (Michael Nygard), High Performance Browser Networking (Ilya Grigorik), and Team Topologies (Skelton & Pais).
cloud-infrastructure
Cloud architecture design for AWS/Azure/GCP, Kubernetes cluster configuration, Terraform infrastructure-as-code, hybrid cloud networking, and multi-cloud cost optimization.
proxmox-infrastructure
Proxmox VE cluster management including VM provisioning, templates, VLAN networking, and CEPH storage.
issue-to-pr
Take a GitHub issue (bare, or tracked on a Project board) from triage to a merge-ready PR through a gated, worktree-isolated pipeline that scales its depth to the task and asks at most one batched question. The gates hold every run: design hardening, tests green, code review clean. Once you approve the PR in-session…