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 nebius/nebius-physical-ai --skill cosmos-data-factory-setupgit clone --depth 1 https://github.com/nebius/nebius-physical-aiWrote 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/nebius/nebius-physical-ai/cosmos-data-factory-setup)<a href="https://agentmods.dev/skills/nebius/nebius-physical-ai/cosmos-data-factory-setup"><img src="https://agentmods.dev/badge/skills/nebius/nebius-physical-ai/cosmos-data-factory-setup/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/nebius/nebius-physical-ai/cosmos-data-factory-setup"><img src="https://agentmods.dev/badge/skills/nebius/nebius-physical-ai/cosmos-data-factory-setup.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 | $0.00040 | $0.00703 |
| Opus 5.5 | $0.00016 | $0.00281 |
| Sonnet 5.5 | $0.00008 | $0.00141 |
| Haiku 4.5 | $0.00004 | $0.00070 |
Grade A, and why
cosmos-data-factory-setup 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 9d 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 — 53 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Cosmos Data Factory setup
Read the shared runbook
before choosing input, image, or GPU settings, then load the IAA or EVG skill.
NVIDIA's reference is Airflow on Kubernetes. NPA translates its DAGs into
npa.workflow/v0.0.1; upstream Helm values, Kubernetes manifests, Airflow API
calls, and payload JSON are not NPA workflow specs. Pinned augmentation templates
and label configs are consumed through the real adapters.
For authoring or review, validate and plan locally, leaving unavailable runtime prerequisites unverified. For a requested run, resolve the authorized project, exact context, staged S3 input, output bucket, and immutable private digests from the current task/configuration. Reuse already supplied values; ask only for missing required choices. Example storage and zero digests are placeholders. IAA/EVG have no automatic demo dataset. Setup does not imply permission to submit, publish images, or destroy infrastructure.
Check S3, Token Factory, HF, NGC, and Nebius authentication and the selected
paidf-iaa/paidf-evg plus labeling access capabilities before allocating GPUs.
Prove private-image pullability on the selected context. Keep credentials in
NPA's store or private environment and forward names through --secret-env;
never expose values in specs, reports, or public handoffs. Runtime access and
image redistribution are separate decisions.
IAA needs the private image-edit and attribute-search wrappers. EVG needs event-video, detection, captioning, visual-qa, and the same attribute-search wrapper. Refer to the compatibility recipes and acceptance evidence linked in the runbook; source inspection does not qualify a rebuilt image. Generation requests one B200 for IAA or two B200s on one node for EVG. Subsequent EVG decoder/labeling stages remain GPU-backed even with a hosted VLM. Use the actual sequential profiles rather than summing Airflow pool reservations.
Follow the runbook's generic configure, validate, plan, submit, status, logs,
artifacts, and cancel commands. Every run preserves reports/upstream.json
and reopens its final handoffs in reports/terminal-validation.json. Report
local validation, deployment readiness, protocol completion, and observed
media/QA quality separately. Cancel only the selected run and verify terminal
state before tearing down owned resources.
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.
- 9d ago First seen · 53 lines · 40 tokens per session scan A b11e25a95f79
cosmos-data-factory-setup is a skill published in the GitHub repository nebius/nebius-physical-ai (50 stars, last pushed today), licensed Apache-2.0. It adds 40 tokens to every session and 703 once invoked, about $0.0002 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 2026-10-13.
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…
gke-workload-security
Audits, configures, and hardens workload-level security controls for Google Kubernetes Engine (GKE) applications and namespaces. Covers running security audits (auditcluster.sh), enforcing Network Policies (default-deny and Dataplane V2 logging), isolating high-risk pods inside GKE Sandbox (gVisor), enforcing Pod…
gke-reliability
Improves GKE workload reliability, using PDBs, health probes, and topology spread constraints. Use when configuring GKE workload reliability, setting up PDBs, or configuring GKE health probes (liveness, readiness, startup). Don't use for generating K8s YAML manifests (use gke-manifest-generation) or disaster recovery…
cloud-architect
Designs cloud architectures, creates migration plans, generates cost optimization recommendations, and produces disaster recovery strategies across AWS, Azure, and GCP. Use when designing cloud architectures, planning migrations, or optimizing multi-cloud deployments. Invoke for Well-Architected Framework, cost…
nemo-automodel-launcher-config
Configure NeMo AutoModel job launches for interactive runs, Slurm clusters, and SkyPilot cloud execution.
azure-mgmt-botservice-dotnet
Azure Resource Manager SDK for Bot Service in .NET. Management plane operations for creating and managing Azure Bot resources, channels (Teams, DirectLine, Slack), and connection settings. Triggers: "Bot Service", "BotResource", "Azure Bot", "DirectLine channel", "Teams channel", "bot management .NET", "create bot".