Borrowing it
Nothing to install: this file belongs to cloudrift-ai/emmy. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/cloudrift-ai/emmy/main/.agents/skills/onboard-model/SKILL.mdgit clone --depth 1 https://github.com/cloudrift-ai/emmyWrote 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/cloudrift-ai/emmy/onboard-model)<a href="https://agentmods.dev/skills/cloudrift-ai/emmy/onboard-model"><img src="https://agentmods.dev/badge/skills/cloudrift-ai/emmy/onboard-model/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/cloudrift-ai/emmy/onboard-model"><img src="https://agentmods.dev/badge/skills/cloudrift-ai/emmy/onboard-model.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00084 | $0.06227 |
| Opus 5.5 | $0.00034 | $0.02491 |
| Sonnet 5.5 | $0.00017 | $0.01245 |
| Haiku 4.5 | $0.00008 | $0.00623 |
Grade A, and why
onboard-model 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 6d 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 — 400 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Onboard or reverify a model
Turn a Hugging Face model ID, an operation mode, and an exact (GPU name, GPU count) into reviewed repository
artifacts:
- a complete, optimized compiler golden under
recipes/<model>/golden/, when the card runs Emmy and coverage is complete; - one recommended serving recipe under
recipes/<model>/recipe.yaml— an Emmy recipe with a verified, prebuiltcloudriftai/vllm-emmy-<model-slug>:<tag>image when Emmy qualifies, otherwise a vLLM or SGLang recipe; - one reusable serving experiment under
experiments/<model>/with a cumulative report and one Git LFS archive per exact GPU platform; recipes/<model>/RESULTS.md: the report beside a valid recipe, or a dated failure record when no recipe qualifies.
The experiment root keeps recipe.yaml, one cumulative RESULTS.md, and results_<gpu-short>x<gpu-count>.tar.gz
per measured platform (<gpu-short> from emmy.hardware.gpu_short_name, e.g. results_rtx4090x1.tar.gz). The
archive holds the timestamped raw run with its system-only experiment records; never commit those records as
top-level files. Track archives with Git LFS. When the caller says LFS is configured, verify the archive's attribute
but do not modify or list .gitattributes. Never commit the ignored dated run directory, loose benchmark output,
plots, compiler run summaries, partial working goldens, or onboarding summaries.
All platforms that share a protocol use one experiment root. A platform run replaces only its own archive and its own
section of the shared RESULTS.md, and removes legacy top-level records only for its platform after verifying the
archive keeps them. Reuse an existing root that represents the protocol, e.g. serving_v100_sxm2_16gb; otherwise
create experiments/<model>/serving/.
Use only the supplied SSH server. The caller owns VM creation and deletion; this skill owns everything on the node and
must tear down every workload before returning. Never switch GPU type, count, provider, quantization, or checkpoint to
rescue a failed run. The node is rented for this run alone and its user has passwordless sudo, so finishing its
provisioning is part of the work: emmy deploy ssh installs Docker and adds the user to the docker group (reconnect
before retrying, because the group applies only to a new login). A missing package, stopped daemon, or missing group
is a node to provision, never a gate failure. Reach the host through Emmy's own commands, not bare-shell probes.
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.
- 6d ago Changed · +1 lines 8d2dacc99cbf
- 14d ago Changed · +2 lines ae28bc374606
- 16d ago Changed · +9 lines fa9ed894891a
- 17d ago Changed · +1 lines 83db685b547a
- 18d ago Changed · +13 lines e5f11c4ddb8e
- 21d ago Changed 34847e58f476
- 22d ago Changed · +1 lines 146bdf57fda9
- 24d ago Changed · -121 lines 759645c31b23
- 26d ago Changed 43055b396113
- 29d ago First seen · 494 lines · 84 tokens per session scan A 2538b4d060e4
onboard-model is a skill published in the GitHub repository cloudrift-ai/emmy (116 stars, last pushed today), licensed Apache-2.0. It adds 84 tokens to every session and 6,227 once invoked, about $0.0003 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 2026-08-30.
Other skills, from other repositories
inference-aiops
Use this skill whenever the user needs to operate a GPU inference cluster — vLLM (OpenAI API + Prometheus /metrics) and Ray Serve / Ray Jobs (Ray dashboard), plus the single-process serving engines SGLang and TGI (Text Generation Inference): a one-shot cluster overview (deployments + total replicas + queue…
sglang-diffusion-performance
Use when choosing the fastest SGLang Diffusion flags for a model, GPU, and VRAM budget.
sglang-diffusion-add-model
Use when adding a new diffusion model or Diffusers pipeline to SGLang.
sglang-diffusion-cache-dit
Workflow for upgrading/integrating cache-dit in SGLang diffusion (multimodalgen): DBCache, DMD calibrator, TaylorSeer, SVDQuant DQ; porting upstream PRs and resolving conflicts against the per-request knob system; adding new cache knobs; building the sglang generate CLI test matrix; precision validation (PSNR / log…
sglang-diffusion-modelopt-quant
Use when quantizing a diffusion DiT with NVIDIA ModelOpt and making the resulting FP8 or NVFP4 checkpoint loadable, verifiable, and benchmarkable in SGLang Diffusion.
hqq-quantization
Half-Quadratic Quantization for LLMs without calibration data. Use when quantizing models to 4/3/2-bit precision without needing calibration datasets, for fast quantization workflows, or when deploying with vLLM or HuggingFace Transformers.