red-hat-data-services

Deploy agents to OpenShift with auto-detected cluster config and refresh MLflow tracking tokens.

Add integration deployment tests (health-check on OpenShift) to any agent in the agentic-starter-kits repo — standard, external-registry, or pre-deployed. Creates conftest.py, test_deployment.py, __init__.py, adds test-integration Makefile target, and updates the CI workflow matrix. Use when implementing integration tests, deployment tests, or health-check tests for a new agent.

Validate whether a new agent template or example belongs in the agentic-starter-kits repo. Two modes: idea mode (interactive questionnaire, no code yet) or existing agent mode (auto-extract from code). Produces a GitHub Discussion draft with fit score and recommendations. Use when proposing a new agent, reviewing an existing contribution's fit, or before writing code for a new template.

Add behavioral testing (pytest + EvalHub) to an agent in the agentic-starter-kits repo. Covers runner compatibility, test files, golden queries, thresholds, EvalHub fixture, Containerfile, docs, and MLflow tracing verification. Use when implementing behavioral tests for a new agent or when the user mentions btest, behavioral tests, eval coverage, or test harness integration.

Adds manual MLflow trace wrapping for tool and agent spans in Level B and C agents where autolog doesn't cover everything.

Researches and classifies a framework's MLflow autolog support level (A, B, or C) to determine what manual tracing is needed.

Creates the tracing.py module with enable_tracing(), health check, and framework-specific autolog configuration.

Orchestrates end-to-end MLflow tracing integration into an agent template, coordinating all sub-skills from research through verification.

Guide for working with Feast (Feature Store) — defining features, configuring feature_store.yaml, retrieving features online/offline, using the CLI, and building RAG retrieval pipelines. Use when the user asks about creating entities, feature views, on-demand feature views, stream feature views, feature services, data sources, feature_store.yaml configuration, feast apply/materialize commands, online or historical feature retrieval, or vector-based document retrieval with Feast.

Internals of the Feast codebase — how each component works, where the key abstractions live, and the data flow through the system. Use when asked how feast apply works, how the registry stores data, how materialization moves data, how get_online_features retrieves features, how the feature server works, how the Kubernetes operator manages deployments, or when navigating the codebase to understand where to make a change.

How to test and debug Feast — running targeted tests, writing unit tests for new components, debugging registry and online store issues, and inspecting live feature store state. Use when writing tests for a new feature, debugging a failing test, investigating a runtime error, or verifying that a change works correctly end-to-end.

Rebase the ODH MLflow fork onto a new upstream MLflow release tag

Review a GitHub pull request and emit a validated local review payload (comments + approval decision)

Resolve PR review comments by fetching unresolved feedback and making necessary code changes

Walk through adding a new container image to the devops-images repo. Guides through creating the Containerfile, updating components.yaml, running generators, and setting up the ServiceAccount.