This skill should be used when developing Helm charts under `charts/*` — templates, `values.yaml`, subchart wrapping, profile overlays, helm-docs annotations, and helm-unittest tests. Covers more than just templates.
Use this skill when working with Materialize's alloy log-processing or metric-processing pipelines — anything under `packages/alloy-pipelines/`. Triggers include: building, modifying, reviewing, or rendering pipelines; working with the embedded JSONSchemas…
This skill should be used when adding, changing, reviewing or vetting an alerting rule or a recording rule: an `alerts:` or `rules:` entry under `packages/queries/`, its severity, capabilities (`requires`) or default-set membership (`enabledByDefault`), the…
This skill should be used when reviewing a pull request or diff. Catches renames and removals on the customer-facing surface that owe a deprecation cycle, and says what is not a breakage.
Use this skill when building, modifying, reviewing, or pushing Grafana dashboards under `packages/dashboards/` (Materialize observability dashboards generated from Rust against the vendored Grafana schemas). Also use it when writing the panel descriptions and…
This skill should be used when investigating a Materialize environment that is unhealthy, slow, lagging, restarting, or refusing connections, and the evidence has to come from the monitoring stack rather than a SQL session. For operators and Materialize…
This skill should be used when writing or editing Markdown prose in this repository, including any page under `docs/content/`, any `SKILL.md`, and `README.md` files. Covers the house voice, the RFC 2119 convention for normative pages, and the authorship…
This skill should be used when changing the Terraform modules under `terraform/`, the E2E test tiers under `test/e2e/`, or the CI that gates them — and when a chart change has consequences for either. Covers how the module and the chart are coupled, what each…
This skill should be used when standing up the monitoring stack against a real or local cluster, or when diagnosing one that is not healthy — pods that will not start, storage that will not attach, config changes that appear to do nothing, or a teardown that…