skills

Use the `ollygarden` CLI to query OllyGarden services, insights, analytics, organizations, and webhooks from the terminal. Use when the user asks to run ollygarden commands, list services or insights, set up or debug webhooks, manage auth contexts (multiple orgs or environments), or pipe OllyGarden data through jq. Triggers on "ollygarden cli", "og cli", "list my services", "fetch insights from cli", "ollygarden auth", "create a webhook", "ollygarden context", "ollygarden --json".

Fetch active OllyGarden service insights from the Olive API and apply remediation fixes to the current codebase. Use when the user asks to get insights, fix insight, address insight, or remediate insight for the current service. Retrieves insights grouped by impact, then applies fixes guided by the API-provided remediation_instructions after user confirmation.

Universal OpenTelemetry auto-instrumentation planning and review guidance. Use when deciding whether agent/library instrumentation is sufficient, when tuning auto-instrumented boundaries, when avoiding duplicate manual spans, or when reviewing auto instrumentation for noise, PII, volume, and business value.

OllyGarden's opinion on when and how to decompose a monolithic OpenTelemetry Collector config into multiple merged files — and when to leave it alone. Use whenever someone wants to split, break up, modularize, or reorganize a large otelcol / collector YAML, or asks "should I split this config", "this collector config is too big to review", "refactor our collector config into multiple files", "organize the collector config by signal / by team". Executes the split directly when warranted (the collector deep-merges repeated --config file: sources), verifies the merged result is behavior-equivalent to the original, and always reports the reasoning — including a deliberate no-op when the config is simple enough not to need it. Plain files and multi-config merge only; not for Terraform/Kustomize/Helm packaging. Layers on facts from otel-collector; hands off behavioral proof to ollygarden-otel-collector-config-validation.

OllyGarden's end-to-end method for validating an OpenTelemetry Collector config by running it — proving a processor or connector transforms, drops, or routes telemetry as intended, not just that the YAML parses. Use whenever someone wants to test, verify, or prove a Collector processor or connector (filter, transform, attributes, redaction, tail_sampling, interval, routing, signaltometrics, count, spanmetrics) against realistic telemetry. Triggers on phrasings like "does my filter actually drop the right spans", "prove this transform works before I ship it", "otelcol validate passes but the rule does the wrong thing", "spin up a throwaway collector and inspect the output". Runs otelcol-contrib validate, then a real collector in Docker or Podman fed by telemetrygen with a file exporter, and asserts the output. Prefer over otel-collector or otel-ottl when the goal is to behaviorally test a config, not look up syntax. Processors and connectors only; receivers/exporters become an OTLP-in / file-out harness.