agentleFS
NetcrackerNetcracker/qubership-ai-agent-telemetry/.claude/skills/qubership-actions-guide/SKILL.md

qubership-actions-guide

Skill1 starsChanged 29 days agoApache-2.0View on GitHub

Navigation-only skill for individual actions in netcracker/qubership-workflow-hub. Use when a workflow needs to consume a specific Qubership action (Docker build/push, version/tag rendering, Maven/npm/Python publishing, package cleanup, Helm charts, security scans, etc.) and you need to find the right action and read its authoritative README. All rules (pinning, permissions, anti-hallucination, naming) live in qubership-workflow-conventions — this skill does not restate them.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 5 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before picking any action, identify what the workflow needs to do and load the relevant supporting document:

    line 10
  2. Distinctive:

    Use the catalog to match each step in the workflow to a Qubership action.

    line 39
  3. Distinctive:

    Ask only what is missing after inferring from context:

    line 60
  4. Very specific:

    Use these exact SHAs in uses: lines.

    line 123
  5. Distinctive:

    Update this table manually when intentionally upgrading to a new release.

    line 124
  6. Distinctive:

    Always resolve the latest release SHA dynamically:

    line 128
  7. Very specific:

    If the GitHub API is unavailable, fall back to the last known SHA: cabbb90e9471163cfac84bd50ff0296b2803b44c (v2.3.0).

    line 140
  8. Very specific:

    When the catalog purpose line is not enough to write the with: block — fetch the action README.

    line 157

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

qubership-actions-guide

Step 1 — identify the operation and load the right guide

Before picking any action, identify what the workflow needs to do and load the relevant supporting document:

OperationLoad
Docker build, push, image migrationdocker.md
Security scan (images, source/deps, k8s cluster)security.md
Email notificationsnotifications.md
Helm chart release, values updatehelm.md
Git tag, GitHub Release, release assetsrelease.md
PR automation (assigners, commit messages)pr.md
CLA / DCO signingcla.md
Maven build, SNAPSHOT deploy, releasemaven.md
npm, Python publishcatalog below — no guide file needed
Cleanup (container images, Maven packages)cleanup.md
APM packages updateutilities.md
Utilities (wait-for-workflow, custom-event, store-input-params)utilities.md
Repository config file resolution without Docker (workflow policy, custom JSON configs)utilities.md
JSON Schema validation (validate a JSON file against a schema, fail the pipeline on violations)utilities.md

Each guide contains: clarifying questions for the user, config file schemas, and pipeline patterns for that domain. Read it before picking actions or asking questions.

Reusable workflows (re-*.yml in netcracker/qubership-workflow-hub) are out of scope — consume them as documented, do not redesign them here.

Step 2 — pick actions from the catalog

Use the catalog to match each step in the workflow to a Qubership action. For full input/output details fetch the action README on demand (see Pin table).

Docker

ActionPurpose
config-resolverSchema-aware config resolver: docker mode by default (same output shape for valid Docker configs, stricter validation), generic flattening for any other schema
docker-config-resolverDeprecated — use config-resolver. Read docker config file, validate, output JSON array for matrix builds
docker-actionBuild & push multi-platform Docker images

Versioning & tagging

ActionPurpose
metadata-actionExtract GitHub context and produce a version string / tags
tag-actionCreate / delete / check Git tags; optional GitHub release creation
branch-actionCreate a new branch from a tag or another branch

Publishing

Ask only what is missing after inferring from context:

  • npm: registry (npmjs / GitHub Packages)
  • Python: target (PyPI / GitHub Packages)
ActionPurpose
maven-releaseMaven artifact release with version bumping and GPG signing
maven-snapshot-deployBuild and deploy Maven SNAPSHOT to Central or GitHub Packages
poetry-publisherBuild, test & publish Python package via Poetry to PyPI

Helm charts

ActionPurpose
chart-versionUpdate version / appVersion in Helm Chart.yaml
charts-values-update-actionUpdate image versions in Helm values files; optionally create release branch

Security & compliance

ActionPurpose
cdxgenGenerate SBOM + CycloneDX vulnerability report from source/deps
k8s-hardening-scanValidate Kubernetes container hardening compliance (Kubescape + Trivy)

Cleanup

ActionPurpose
container-package-cleanupRemove stale container or Maven package versions from registry

PR & collaboration

ActionPurpose
cla-assistantCLA / DCO signing via PR comments
pr-assignerAuto-assign reviewers based on config / CODEOWNERS
pr-add-messagesAppend commit messages to PR description

Notifications

ActionPurpose
email-actionSend email notifications via SMTP with fallback to repository variables for connection settings

Utilities

ActionPurpose
config-resolverResolve any JSON config file to flat JSON (also listed under Docker) — generic mode for non-Docker configs, e.g. workflow policy
apm-packages-updateRun apm update --yes and open a PR with the changes in the current repo
ghcr-discover-repo-packagesDiscover all GHCR packages for a repo — feeds security scan, cleanup, or any step needing the image list
custom-eventEmit repository_dispatch event with JSON payload
smart-downloadDownload workflow artifacts by name, IDs, or glob pattern
store-input-paramsPersist workflow_dispatch inputs as artifact
wait-for-workflowWait for a specific GitHub Actions workflow run to complete
verify-jsonValidate a JSON file against a JSON Schema
assets-actionUpload files/dirs to a GitHub release, auto-archives directories

Deprecated (do not use): commit-and-push, pom-updater, tag-checker, archive-and-upload-assets.

Pin table

Use these exact SHAs in uses: lines. For SHA pinning rules see qubership-workflow-conventions → Pinning. Update this table manually when intentionally upgrading to a new release.

netcracker/qubership-workflow-hub

Always resolve the latest release SHA dynamically:

1. WebFetch → https://api.github.com/repos/netcracker/qubership-workflow-hub/releases/latest
   → extract .tag_name (e.g. v2.3.0)
2. WebFetch → https://api.github.com/repos/netcracker/qubership-workflow-hub/git/ref/tags/<tag_name>
   → extract .object.sha and .object.type
   - if .object.type == "commit" → use .object.sha directly
   - if .object.type == "tag" (annotated tag) → WebFetch .object.url → extract .object.sha (the commit SHA)
3. Use that commit SHA in uses: lines. Never use .target_commitish — it is a branch name, not a SHA.

If the GitHub API is unavailable, fall back to the last known SHA: cabbb90e9471163cfac84bd50ff0296b2803b44c (v2.3.0).

uses: netcracker/qubership-workflow-hub/actions/docker-action@<resolved-sha>  # <resolved-tag>

Third-party actions (update manually on upgrade)

RepoLatest tagSHA
netcracker/release-drafterv1.0.086f4276a3894b5af70480e826c32fe3648ac6a70
actions/checkoutv6.0.3df4cb1c069e1874edd31b4311f1884172cec0e10
actions/upload-artifactv7.0.1043fb46d1a93c77aae656e7c1c64a875d1fc6a0a
actions/download-artifactv8.0.13e5f45b2cfb9172054b4087a40e8e0b5a5461e7c
actions/setup-pythonv6.2.0a309ff8b426b58ec0e2a45f0f869d46889d02405
actions/create-github-app-tokenv3.2.0bcd2ba49218906704ab6c1aa796996da409d3eb1

When the catalog purpose line is not enough to write the with: block — fetch the action README. Never write inputs from memory.

1. WebFetch → https://api.github.com/repos/netcracker/qubership-workflow-hub/releases/latest
   → extract .tag_name
2. Resolve commit SHA via /git/ref/tags/<tag_name> (see Pin table above for the exact procedure)
3. WebFetch → https://raw.githubusercontent.com/netcracker/qubership-workflow-hub/<commit-sha>/actions/<name>/README.md

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.