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 secure-image-buildgit 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/secure-image-build)<a href="https://agentmods.dev/skills/nebius/nebius-physical-ai/secure-image-build"><img src="https://agentmods.dev/badge/skills/nebius/nebius-physical-ai/secure-image-build/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/secure-image-build"><img src="https://agentmods.dev/badge/skills/nebius/nebius-physical-ai/secure-image-build.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 | $0.00057 | $0.01165 |
| Opus 5 | $0.00028 | $0.00583 |
| Sonnet 5 | $0.00011 | $0.00233 |
| Haiku 4.5 | $0.00006 | $0.00117 |
Grade A, and why
secure-image-build 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 2d 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 — 96 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Secure Image Build
Treat every official development push as public and irreversible. Use one
official namespace: ghcr.io/nebius/nebius-physical-ai/<image>. Tag reviewed
development bytes dev-<full-git-sha> and promote only that exact digest to the
supported release tag.
Load The Governing Procedures
Read and follow these before acting; do not duplicate their detailed commands:
skills/atomic/solution-licensing/SKILL.mdskills/atomic/build-and-push-image/SKILL.mdskills/workflows/contribute-workbench-image/SKILL.mdskills/atomic/third-party-eula-preflight/SKILL.mdskills/atomic/testing-conventions/SKILL.mddocs/workbench/container-packaging.md
For live validation also load skills/atomic/gpu-selection/SKILL.md,
skills/atomic/submit-workflow/SKILL.md, and
skills/atomic/protect-nebius-infra-details/SKILL.md.
Mandatory Sequence
- Resolve the exact checked-out commit. Require a full 40-character Git SHA
and the immutable tag
dev-<full-git-sha>; refuse moving or abbreviated development tags. - Require
redistribution: publicinnpa/docker/workbench/packaging-contract.yamland check the current restriction inventories innpa/src/npa/deploy/images.py. Hard-refuserestrictedimages from every official GHCR tag; resolve any disagreement between those sources before publication. Read current eligibility and quarantine state from those sources, rather than treating a historical product example as a fixed inventory. The supportedcosmos3-servingrelease is bound to its accepted, payload-scanned and GPU-validated digest; historical restricted bytes remain prohibited. Future development bytes must pass the sequence below. - Before any public push, run the repository packaging/license guards and inspect the locally built artifact, including layers, history, and OCI config. Refuse credentials, secrets, customer data, live infrastructure identifiers, gated weights/data, proprietary SDK/runtime payloads, or cached EULA acceptance anywhere in the artifact.
- Require policy-approved, digest-pinned bases/dependencies or an existing documented packaging exception. Enforce the non-root runtime and packaging contract. When SkyPilot needs the bootstrap contract, prove the behavior and exact OCI attestation; never add the label without the proof.
- Produce the SBOM and run the repository-supported vulnerability, secret, license, payload, revision/provenance, and bootstrap checks. Do not push the public development tag until every pre-publication gate passes.
- Push with explicit
packages: write, then resolve the tag once to its OCI digest. Verify revision/provenance and SBOM attestations, rerun exact-digest payload/security checks, and prove an anonymous pull without ambient auth. - Run the real functional workflow on a compatible physical GPU using the immutable development digest. An import, startup, or CUDA-availability check is not functional evidence.
- Promote only the validated digest to the supported public release tag through
npa.deploy.publish_public. Verify anonymous pullability and exact digest identity independently after promotion; record the exact digest identity inpublic_release_manifest.json. Scheduled health must compare the release tag anonymously with thatpublished_digest, without depending on dev-tag retention. - On failure, delete only the exact run-owned development version after matching package, tag, and digest. Never infer ownership from a name or delete a shared/release digest. Record that deletion cannot revoke prior downloads. Retain a successful dev tag only when the documented release policy requires it to preserve the release's shared digest/provenance.
What ships with it
1 file 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.
- 2d ago Changed · +21 lines 2ef38b4f4bdb
- 4d ago Changed · +3 lines b64c182b8aeb
- 10d ago First seen · 72 lines · 57 tokens per session scan A 3bead60e666e
secure-image-build is a skill published in the GitHub repository nebius/nebius-physical-ai (27 stars, last pushed today), licensed Apache-2.0. It adds 57 tokens to every session and 1,165 once invoked, about $0.0003 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 2026-08-30.
Other skills, from other repositories
deploy-docker-compose
Run the Omnigent server as a Docker compose stack (server + Postgres) on any Docker host — your laptop, a VPS, EC2 by hand, or as the base layer of any container-platform deploy. Invoke when the user wants to build the image, bring up the compose stack, debug the stack on a host they already have, or extend the stack…
compute-env-setup
Set up a reproducible Feynman compute environment for research jobs. Use when a task needs Python/R packages, GPU libraries, containers, Modal, SSH, caches, or managed model runtime setup.
securing-kubernetes-on-cloud
This skill covers hardening managed Kubernetes clusters on EKS, AKS, and GKE by implementing Pod Security Standards, network policies, workload identity, RBAC scoping, image admission controls, and runtime security monitoring. It addresses cloud-specific security features including IRSA for EKS, Workload Identity for…
detecting-privilege-escalation-in-kubernetes-pods
Detect and prevent privilege escalation in Kubernetes pods by monitoring security contexts, capabilities, and syscall patterns with Falco and OPA policies.
implementing-rbac-hardening-for-kubernetes
Harden Kubernetes Role-Based Access Control by implementing least-privilege policies, auditing role bindings, eliminating cluster-admin sprawl, and integrating external identity providers.
docker-socket-mount
Docker / containerd socket mounted into a container → host RCE. Common in CI runners, GitOps controllers (ArgoCD, Flux), and 'Docker-in-Docker' setups. Single-command escape via docker run --rm --privileged -v /:/host alpine chroot /host.