performance

name	performance
description	Evaluate and improve performance by identifying hot paths, reducing unnecessary allocations/work, and optimizing data flow and I/O behavior without changing required functionality. Use when addressing latency, throughput, CPU/memory overhead, startup cost, or scaling bottlenecks.

Performance Workflow

Inputs

• scope: files, module, package, service, or subsystem.
• mode (optional): review-only or optimize (default).
• target_metric (optional): latency, throughput, CPU, memory, startup, or mixed.

References

Use selective loading:

• Always start with references/philosophy/overview.md.
• Load only active-language checklists:
• references/checklists/go.md
• references/checklists/javascript-typescript.md
• references/checklists/python.md
• references/checklists/rust.md
• In polyglot changes, load only the checklists for touched languages.

Procedure

1. Identify measurable target metrics and likely hot paths.
2. Inspect algorithmic complexity, allocation patterns, I/O, and contention.
3. Propose optimizations with expected tradeoffs.
4. In optimize mode, implement focused changes with minimal behavioral risk.
5. Validate with benchmarks/profiling/tests where available.
6. Summarize gains, tradeoffs, and follow-up opportunities.

Performance Focus Areas

• Algorithmic complexity and unnecessary work.
• Allocation frequency and object lifetime.
• I/O batching/caching and round-trip reduction.
• Concurrency/parallelism overhead and lock contention.
• Data structure choices for access patterns.

Output Contract

1. Performance findings (bottleneck + evidence).
2. Optimization summary (when mode=optimize).
3. Measurement/validation results and residual tradeoffs.