publish-helm
This skill publishes Helm charts to OCI registries end to end, handling chart validation, PR creation, merging, tagging, and verification.
Identity proven by GitHub's verified organisation badge (nebius)
- Agents
- Claude Code
- License
- Apache-2.0 (permissive)
- Pin
- b1924depinned Oct 9, 2026
- Context cost
- about 2,000 tokenswhen the skill loads
- GitHub stars
- 7on nebius/nebius-ps-services
- Scripts
- 1 script
What it does
“Use only when explicitly asked to publish Helm charts end to end: collect chart/OCI inputs, set up optional assets, prepare/merge a PR, tag, wait, verify the OCI chart, and report. Also supports setup-only guidance.”
Generated summary from the skill's own text
This skill publishes Helm charts to OCI registries end to end, handling chart validation, PR creation, merging, tagging, and verification. It supports four modes: setup-only asset installation, prep for PR review, publish for execution, and complete for full automation. The skill manages the entire release workflow with guardrails against unsafe operations and requires explicit version and destination inputs.
When to use it
- Publishing Helm charts to OCI registries with full workflow automation
- Setting up release assets for projects without existing publish flows
- Running individual phases like prep, publish, or tag creation independently
- Verifying published chart artifacts and digests after release
When not to
- When publishing to non-OCI registries or traditional Helm repositories
- If chart version and destination cannot be explicitly specified by the user
- When agentic SDLC workflows or protected branch policies cannot be observed
Install
These commands install from the publisher's repository and need no account.
skills CLI
npx skills add nebius/nebius-ps-services --skill publish-helmInstalls from the repository's default branch, which may have changed since the commit we checked.
GitHub CLI 2.90 or later
gh skill install nebius/nebius-ps-services publish-helm@b1924de70353e4e688be5017ec608ec9b89c8ff4Installs the commit we checked (b1924de).
Copy by hand
Download the folder at b1924de and put it in .claude/skills/publish-helm for Claude Code and in .agents/skills/publish-helm for Codex, Cursor, Gemini CLI, GitHub Copilot and OpenCode.
Copy for your agent
A prompt for any coding agent. It installs this skill from GitHub at the pinned commit, copies the files as they are, and runs nothing unless you ask.
Prompt
Install the agent skill "publish-helm" by Nebius. Source, pinned to commit b1924de70353e4e688be5017ec608ec9b89c8ff4: https://github.com/nebius/nebius-ps-services/tree/b1924de70353e4e688be5017ec608ec9b89c8ff4/skills/publish-helm 1. Download that folder at exactly this commit. Do not use the default branch. 2. Copy every file unchanged, keeping the folder structure, into .claude/skills/publish-helm/ for Claude Code and .agents/skills/publish-helm/ for Codex, Cursor, Gemini CLI, GitHub Copilot and OpenCode. Create the folders if they do not exist. 3. Do not edit the files, and do not run any script in the folder unless I ask you to. 4. When you are done, list the files you copied and where you put them. Listing and checks: https://www.agentnative.dev/skills/nebius/publish-helm
SKILL.md at the pin
--- name: publish-helm description: "Use only when explicitly asked to publish Helm charts end to end: collect chart/OCI inputs, set up optional assets, prepare/merge a PR, tag, wait, verify the OCI chart, and report. Also supports setup-only guidance." disable-model-invocation: true --- # Publish Helm ## Help For `$publish-helm --help` or `$publish-helm -h` (including native Claude forms), return concise help and stop before any workflow step. State the purpose and invocation policy. Show exact usage for every public action. Describe each public action, positional argument, and flag in one concise line, including `-h, --help`; say "No additional public flags" when there are no others. Use only the documented public interface. For internal or coordinator-only skills, state that boundary and that no standalone public workflow action exists. After the selected `SKILL.md` is loaded, help is report-only: do not call any additional tools, inspect project state, or modify files, private state, Git, or external systems. Never expose private helper actions or flags or treat help as workflow authorization. ## Agent Compatibility Use `$publish-helm` in Codex, `/publish-helm` in Claude Code, or `/skills:publish-helm` in the Claude plugin. Dollar-prefixed skill examples refer to the same named skill on either host; use the native invocation syntax. Preserve the declared invocation policy, approvals and workflow ownership. Use available native tools; an unavailable required capability is a blocker, never permission to bypass a guard or claim unobserved behavior. ## Purpose Publish a Helm chart release from the current project folder. This is a doer-first skill: setup/guidance is still supported, but release execution is the primary workflow when the user asks to publish. ## Use This Skill For - Publishing a Helm chart to an OCI registry end to end. - Setting up chart release assets only when requested or missing. - Running one explicit phase: `setup`, `prep`, `publish`, or `complete`. - Producing a final publish report with PR, tag, workflow, chart ref, and pull verification. ## Inputs Accepted Common flags: - `--mode setup|prep|publish|complete`; use `complete` for an end-to-end publish request. - `--tag X.Y.Z[-prerelease]` or `<tag-prefix>-vX.Y.Z[-prerelease]`. - `--project-dir <path>`; default current working directory. - `--main-branch <branch>`; default the repository default branch. - `--tag-prefix <prefix>`; derive from chart name only when unambiguous. - `--merge-method squash|merge|rebase`; default `squash`. - `--wait` or `--no-wait`; default `--wait` for `publish` and `complete`. Helm inputs: - `--chart-dir` - `--chart-name` - `--oci-repository`, such as `oci://registry.example.com/org/charts` - `--public-verify` when anonymous pull verification is expected - optional extra lint or template smoke arguments `--oci-repository` is the repository base. It must not include the chart basename or chart version tag; Helm infers those from chart metadata. ## Required Release Inputs Before running `prep`, `publish`, or `complete`, resolve these from explicit user input or unambiguous project configuration: - Release version/tag is required. If the user does not provide it, ask for `X.Y.Z[-prerelease]` or the full `<tag-prefix>-vX.Y.Z[-prerelease]`; do not infer it from `Chart.yaml`, the latest Git tag, branch names, or changelog text. - Publish destination is required. If the current project does not already have a chart publish workflow with a configured target and the user did not provide a destination, ask for the target registry or OCI repository details. - For generic Helm CLI publishing and local verification, use an OCI repository base such as `oci://registry.example.com/org/charts`. If the user provides a full chart reference ending in the chart name, treat it as the report/pull reference and pass only the repository base to helper scripts. - For project workflows that derive the upload target from provider-specific variables such as region and registry ID, inspect the workflow and use those variable names as the contract. Do not hardcode concrete registry IDs, registry URLs, project IDs, endpoints, or secret values in reusable skill sources. ## Destination Forms - Generic Helm form: `helm push` receives the OCI repository base without chart name or version, and `helm pull` appends `<chart-name> --version <version>`. - Workflow form: a project workflow may publish from environment, GitHub variables, or secrets and derive the final OCI reference at runtime. `--oci-repository` does not override such a workflow unless the workflow explicitly supports it; use it for local final verification and reporting. ## Workflow Read [ordinary publication](references/ordinary-publication.md) before effects. 1. Resolve required explicit version, project, destination, tag prefix and live default. Setup-only installs requested assets and returns without publication. 2. For complete/prep, use create-pr intake before branch movement. Reuse the current feature; from default create a feature carrying the local work. 3. The helper's prep mode changes content only. Validate all repository changes, commit and push through create-pr's canonical whole-repository transaction. 4. Run local review-pr, repair safe findings through the private preparation handoff, revalidate and re-review. Stop for unsafe findings or human objections. Public prep stops at the reviewed PR. Complete routes merge only to merge-pr. 5. Observe actual protected merge, prove its exact result is in remote default history, and require passing applicable result CI. A queue is still pending. 6. In a clean isolated clone at that exact result, verify release metadata. Use helper tag mode to create an annotated tag, record its object, then push mode with that exact object. Never tag latest main or alter the source worktree. 7. Observe the exact tag/result publication workflow and verify the published artifact and digest. Report all evidence or the precise blocker. ## Setup Assets Use setup mode when the chart does not already have a release flow: - `assets/CHANGELOG.md.template` - `assets/publish-helm.sh.template` The project-local helper script is optional, but it is a maintained runnable helper template, not a documentation stub. Keep it behaviorally aligned with the skill-owned `scripts/publish-helm-doer.sh`, which remains the canonical doer path. ## Guardrails - Agentic SDLC is excluded; preserve its owners and exact promoted SHA. - Complete authorization covers the necessary ordinary PR/review/merge/tag flow; do not request repeated commit, merge or routine Actions approval. - Missing inputs, credentials, unsafe findings, human objections and additional protection requirements are blockers. Never bypass protections or use a PAT merge fallback. Agent review stays local; Actions runs deterministic jobs only. - Prep never stages, commits, switches branches or pushes. create-pr owns those effects with normal hooks and repository-root `git add -A`. - Publishing tags never edits release content. The reviewed changelog and chart metadata (when applicable) must already be present in the verified result. - Never force-push, replace a tag, tag an unrelated default tip, or claim queued work is merged. Preserve unrelated user work and exact resume evidence. ## Learning Loop When using this skill, capture durable, reusable, public-safe learnings in the narrowest appropriate surface only when the task contract allows source edits. For read-only/report-only work, or when a learning is not public-safe, evidence-backed, in scope, or free of unverified/vendor-specific claims, do not edit skill sources; report that it was skipped. Do not capture secrets, private URLs, customer data, raw logs, or one-off local state. ## Output Contract Return: - mode, chart directory, chart name, tag, version, tag prefix - release branch, PR URL, and merge result when `complete` mode is used - pushed tag and workflow run URL/conclusion - OCI chart reference and pull verification result - validation commands run - blockers, skipped live checks, or required user approvals ## Resources - `scripts/publish-helm-doer.sh` - `assets/CHANGELOG.md.template` - `assets/publish-helm.sh.template`
From nebius/nebius-ps-services at b1924de, shown as plain text under its license: Apache-2.0 (permissive). The license text is in LICENSE.
View SKILL.md on GitHub at b1924deFiles
- SKILL.md8.4 KB · b112b5b0b33a
- README.md3.6 KB · 31770eddf832
- agents/
- openai.yaml492 B · bf6e5083b8b7
- assets/
- CHANGELOG.md.template912 B · f8e2ff48cf22
- publish-helm.sh.template22 KB · 67d96054da6a
- evals/
- evals.json3.2 KB · f8a48adc382f
- trigger-prompts.csv883 B · 24a1ba045172
- references/
- ordinary-publication.md2.7 KB · 6c4ed04c49f3
- scripts/
- publish-helm-doer.sh22 KB · 7b24b2875b5e
9 files, 64 KB in total. Each file has a sha256 in the pin's manifest; the first 12 characters are shown.
Skill folder on GitHub at b1924deMore from Nebius
agent-development
This skill guides building the NPA chat agent backend using a grounded-first routing architecture that minimizes token spending.
Claude Code, Codex, Cursor +4 · about 5.7K tokens · pinned 7 days ago
author-npa-workflow
Helps authors write and validate NPA workflow YAML specifications, which are declarative state machines that orchestrate tools via SkyPilot.
Claude Code, Codex, Cursor +4 · about 3.3K tokens · pinned 7 days ago
cosmos
This skill enables the agent to work with Cosmos, a world model for synthetic data and video generation.
Claude Code, Codex, Cursor +4 · about 1.6K tokens · pinned 7 days ago
cosmos3-npa-workflow
This skill helps agents write, validate, and submit npa.workflow YAML specifications for Cosmos 3 image and video generation tasks.
Claude Code, Codex, Cursor +4 · about 1.7K tokens · pinned 7 days ago
cosmos3-post-training
This skill guides planning and understanding of Cosmos3 supervised fine-tuning workflows for robotic policy training on NPA infrastructure.
Claude Code, Codex, Cursor +4 · about 2.5K tokens · pinned 7 days ago
curobo
This skill enables agents to run NVIDIA cuRobo V2 motion planning for robotic arm pose planning and benchmarking on CUDA GPUs.
Claude Code, Codex, Cursor +4 · about 1.5K tokens · pinned 7 days ago
Related in Cloud and DevOps
kubernetes-skill
This skill helps agents diagnose and fix six common Kubernetes failure modes: insecure workload defaults, resource starvation, network exposure, privilege…
by LukasNiessenCommunity
Claude Code, Codex, Cursor +4 · about 1.1K tokens · pinned 7 days ago
n8n-cli
This skill enables agents to manage n8n workflows, credentials, executions, and instance resources via the n8n CLI command-line tool.
by n8nOfficial publisher
Claude Code, Codex, Cursor +4 · about 2.9K tokens · pinned 7 days ago
managing-streamlit-apps
Manage Streamlit apps running in PostHog sandboxes through the full lifecycle: create, set or edit source code, start and stop the sandbox, poll status, and…
by PostHogOfficial publisher
Claude Code, Codex, Cursor +4 · about 1.4K tokens · pinned 7 days ago
release-candidate-prep
Prepares a local Python SDK release candidate in a dedicated Git worktree when explicitly invoked with a version number.
by OpenAIOfficial publisher
Contains scripts
Claude Code, Codex, Cursor +3 · about 4.8K tokens · pinned 7 days ago
cloud-run-basics
Deploys and manages Google Cloud Run services, jobs, and worker pools for running containerized applications or source code.
by GoogleOfficial publisher
Claude Code, Codex, Cursor +4 · about 3.9K tokens · pinned 7 days ago
debug-openshell-cluster
Diagnoses OpenShell gateway deployment problems across Docker, Podman, Kubernetes, and VM compute platforms.
by NVIDIAOfficial publisher
Claude Code, Codex, Cursor +4 · about 19.1K tokens · pinned 7 days ago
Open roles at Nebius
5 of 75 open roles on the jobs board
- Senior Software Engineer (Capacity and Quota Management)
Amsterdam, Netherlands; Berlin, Germany; United Kingdom
- Senior Backend Software Engineer (Observability)
Amsterdam, Netherlands (Remote)
- Lead Security Automation Engineer
Israel
- Senior Software Engineer (Data Platform, C++)
London, United Kingdom
- Senior AI Infrastructure Systems Engineer (R&D / GPU / AI)
United States