aiops-skills¶
Component onboarding automation for ODH and RHOAI on the Konflux CI/build platform. Provides an interactive create-then-validate pipeline: collect onboarding parameters through guided Q&A, generate a validated component_onboarding_details.yaml, create or update Jira tickets with the YAML attached, and validate existing onboarding tickets against a JSON Schema with cross-checks for branch naming, Dockerfile digest pinning, and product-specific requirements.
Supports both ODH (CI and Release builds with optional release tags) and RHOAI (architecture selection, release categories, target version-derived branch naming, and operator manifest handling). All Jira operations use deterministic Python scripts run via uv with Atlassian REST API authentication (JIRA_USER_EMAIL + JIRA_API_TOKEN). When no Jira URL is provided, the create skill automatically clones a product-specific template ticket (ODH: RHOAIENG-35683, RHOAI: RHOAIENG-17225).
Plugin Details
- Version: 0.1.0
- Author: opendatahub-io
- License: Apache-2.0
- Category: DevOps & CI/CD
- Repository: opendatahub-io/aiops-infra
- Tags: devops testops odh rhoai konflux onboarding ci-cd release automation
Pipeline¶
Skills¶
| Skill | Description | Invocable |
|---|---|---|
/create-component-onboarding-jira |
Interactively collect component onboarding parameters and create/update a Jira ticket | |
/validate-component-onboarding-jira |
Pre-flight validation for ODH component onboarding — fetches Jira, downloads YAML, validates against schema |
Installation¶
Claude Code
/plugin install aiops-skills@opendatahub-skills
OpenAI Codex — add the marketplace, then enable aiops-skills from the /plugins browser:
codex plugin marketplace add opendatahub-io/skills-registry
Architecture¶
Two skills form a create-then-validate pipeline for component onboarding:
create-component-onboarding-jira runs an 8-step interactive flow: parse input and check prerequisites (uv, jq, Jira credentials), set up a working directory, optionally fetch existing Jira details, collect parameters via guided Q&A with product-aware conditional logic (ODH vs RHOAI), generate YAML via generate_onboarding_yaml.py, validate against JSON Schema via validate_yaml_schema.py, check Dockerfile digest pinning for RHOAI via check_dockerfile_digests.py, create or update a Jira ticket (cloning product-specific templates if no URL provided) via update_jira_issue.py, and report results.
validate-component-onboarding-jira runs a 6-step validation flow: check prerequisites, create working directory from Jira URL via init_workdir.sh, fetch issue details via fetch_jira_details.py, download YAML attachment via download_jira_attachment.py, validate against schema with cross-checks (branch naming for RHOAI, Dockerfile digest pinning), and update Jira with validation-successful or validation-failed labels via update_jira_issue.py.
Both skills use deterministic Python scripts (run via uv) for all Jira API operations and YAML schema validation. The component_onboarding_details.schema.json defines conditional requirements based on product_context (ODH vs RHOAI) and is_operator fields. Shared scripts: init_workdir.sh (working directory setup), check_prerequisites.sh (tool and env var checks), fetch_jira_details.py, update_jira_issue.py, validate_yaml_schema.py, check_dockerfile_digests.py.