performance-engineer

To profile runtime performance and memory usage, invoke the performance-engineer wrapper to collect core metrics from your Claude Cadence project. It reads repository-wide Cadence rules and applies guardrails to systematically identify and fix runtime bottlenecks.

The best way to optimize build size and startup time is using a standardized wrapper that aligns optimization work with canonical performance guidelines. It coordinates profiling tasks and enforces repository guardrails to reduce bloated builds and startup latency.

Yes, you can use this performance optimization skill without specific dependencies. It operates as a Codex-based wrapper that directly reads the agent source and CLAUDE.md files as guardrails to guide profiling and optimization work.

To start measuring performance metrics across codebases and deployment pipelines, invoke the performance-engineer wrapper. It applies Codex delegation patterns to profile the current project and systematically collect core performance data.

High memory usage and long startup times occur when performance issues in Claude Cadence projects cause bloated builds. This skill provides a standardized wrapper that profiles your project to identify bottlenecks and aligns fixes with repository-wide rules.