Micro-benchmark and iteratively improve hot paths in any codebase — baseline first, mandatory improvement suggestions (bottleneck + 3 creative ideas + recommended next step), bounded attempts, user lock-in checkpoints, then human WHY comments after lock-in. Use when the user asks to benchmark, profile, optimize performance, compare implementations, or measure µs/op before shipping perf changes. Applies to Bun/Node, React, Next.js, browser code, and extensions.

日本語の概要は準備中です。原文の説明を表示しています。

Cleanup code, optimize, refactor, delete the 80% to keep the 20%

日本語の概要は準備中です。原文の説明を表示しています。

Build, run, or update minimal evaluation benchmarks for AI prompts, LLM calls, multi-model comparisons, and core functions. Use when creating new eval scripts, comparing models side-by-side, filtering specific candidate models, running benchmarks, or inspecting eval.md output artifacts.

日本語の概要は準備中です。原文の説明を表示しています。

Generates browser extension brand icons, Chrome Web Store promo tiles, promotional screenshots, brand guides, and listing copy. Uses OpenRouter gpt-image-2 via lib/blog-image-generator. Use when creating extension branding, Chrome Web Store assets, app icon concepts, store descriptions, or /extension-store-assets workflows.

日本語の概要は準備中です。原文の説明を表示しています。

Optimizes code, queries, and features based on actual usage by removing unused fields, aggregations, JOINs, and unused complexity. Use when the user invokes /optimize or asks to slim queries to what the UI actually uses.

日本語の概要は準備中です。原文の説明を表示しています。