メニュー
Use for all XppAI AX 2009 tasks as the canonical core baseline (architecture, language, lifecycle, hard constraints, and evidence labels).
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Use when analyzing X++ AX 2009 profiler traces, code stack traces, or execution paths to identify performance bottlenecks — repeated call chains, quadratic loops, expensive constructor+calc combos, excessive tax/totals recalculation, or deep framework cascades.
Use when reviewing X++ or general ERP source code for architectural weaknesses, design gaps, fragile sequencing, responsibility mismatches, technical debt, or structural risks — not just syntax or performance issues.
Use when given any X++ AX 2009 artifact — stack trace, method, class, form XPO, posting code, or table — and you need a full structured multi-skill analysis applied automatically based on artifact type.
Use when proposing X++ AX 2009 code fixes for identified performance bottlenecks or behavioral bugs — minimal, production-safe changes that preserve business logic and fit ERP team review standards.
Use when deep AX 2009 analysis is required (performance, posting, totals/tax, framework patterns, customization safety, and architecture judgment).
Use when reading unfamiliar X++ AX 2009 code — methods, classes, forms, tables, or XPO extracts — to understand what it does, what triggers it, what it calls, and where it fits in the AX execution model.
| name | xppai-core |
| description | Use for all XppAI AX 2009 tasks as the canonical core baseline (architecture, language, lifecycle, hard constraints, and evidence labels). |
Canonical shared baseline for all XppAI skills.
Use these labels consistently across all skills:
ConfirmedLikelyHypothesisUnknown<GBR>, <GIN>, <GJP>, <GSA>, <GTH>TextIo has no .eof(); use status() == IO_Status::Okselect count(RecId) result is read from buffer RecId