원클릭으로
chessbeast
chessbeast에는 rlefko에서 수집한 skills 3개가 있으며, 저장소 수준 직업 범위와 사이트 내 skill 상세 페이지를 제공합니다.
이 저장소의 skills
Chess-correctness review for ChessBeast analysis changes. Use when a diff touches packages/core/src/classifier, packages/core/src/exploration, packages/core/src/storage, packages/pgn/src, packages/llm/src/explorer, packages/cli/src/orchestrator, or services/. Also use when the user says things like "chess correctness", "check the notation change", "review the classifier change", or "are these plies right". Reads docs/analysis-conventions.md as the source of truth.
Review annotation output quality against the annotation conventions. Use when a change touches packages/llm/src/narration, packages/llm/src/annotation, packages/llm/src/themes, or packages/pgn/src/transformer, when prompt text changes anywhere in the repo, or when reviewing a sample annotated PGN. Trigger phrasings include "review the narrator change", "check these prompt edits", "does this comment output look right", "review annotation quality", "check the density change", or "read this annotated game". Reads docs/annotation-conventions.md as the source of truth.
Performance review for ChessBeast pipeline changes. Use when a diff touches packages/core/src/exploration, packages/core/src/storage, packages/llm/src/narration/narrator.ts, packages/llm/src/client, packages/cli/src/orchestrator, packages/debug-gui/src/server, or a pool.py under services/. Also use when the user says things like "perf review this", "did we regress throughput", "is this loop too slow", "check the engine budget", "review LLM fan-out", or "is the GUI slowing the pipeline". Reads docs/performance-conventions.md as the source of truth.