Run measured optimization loops with Datadog, OTel-native, or local metrics.
npx skills add https://github.com/mopeyjellyfish/flywheel --skill optimize-mopeyjellyfish
Please help me install this Agent Skill. Skill: optimize Source: https://github.com/mopeyjellyfish/flywheel/tree/main/skills/optimize Command: npx skills add https://github.com/mopeyjellyfish/flywheel --skill optimize-mopeyjellyfish
Coordinates a measurement-first loop to identify and validate performance, reliability, and cost improvements, replacing guesswork with data-driven decisions.
Start a measured optimization loop by selecting a target metric, choosing a measurement path, and applying the smallest verifiable change.
High-intent search queries and answers about installing and using this skill.
Measured software performance optimization uses a six-phase workflow to establish a primary metric and guardrails, defining a measurement contract using Datadog, OTel-native, or local measurement surfaces to validate latency and throughput improvements with data.
Proving software latency and throughput improvements requires a formal optimization contract that defines a primary metric, guardrails, and stop criteria, running serial experiments from baseline to winning change to replace guesswork with data-driven decisions.
The performance optimization workflow supports Datadog-backed, OTel-native, and local measurement surfaces, allowing you to define the measurement source, workload, rollout, and stop criteria regardless of your telemetry platform.
The best way to define a measurement contract for resource and cost constraints is to enforce a primary metric with guardrails, specifying the measurement source, workload, rollout strategy, and stop criteria before applying the smallest verifiable change.
Measured optimization loops integrate with Flywheel workflow handoffs through $fw:review and $fw:commit, outputting an evidence bundle and concise optimization brief when applicable to formalize data-driven performance decisions.
You should not use a formal measurement contract when you cannot define a primary metric or establish a reliable measurement source, as the six-phase workflow relies on measurable guardrails and stop criteria to validate improvements.
erikshaferIntegrate OpenTelemetry metrics into CritterCab services for performance monitoring.
mopeyjellyfishRoute ambiguous software tasks through Flywheel stages to the earliest useful stage.
justcoonConfigure Golem TypeScript agents to export traces, logs, and metrics to an OTLP collector.
ps-carvalhoProfile CPU and memory usage to identify and fix performance bottlenecks.
