メニュー
Analyze page content shape and extract layout decision features (deterministic, no LLM)
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
| name | content-analyzer |
| description | Analyze page content shape and extract layout decision features (deterministic, no LLM) |
| version | 4.0.0 |
| domain | office |
| layer | understanding |
| tags | ["slides","content","analysis","understanding"] |
| taskTypes | ["slides-generation"] |
| priority | 50 |
| author | genesis-ai |
| source | local |
| execution-mode | provider |
This skill is implemented as a NestJS Provider (code-based tool). This SKILL.md provides metadata only -- execution is handled by the corresponding .skill.ts file.
Deterministic content analysis skill that extracts layout decision features from page content. Replaces hardcoded template capacity configuration with data-driven content shape analysis.
ContentAnalysisResult with section types, content metrics, logical structures, data density, visual complexity, recommended layout, grid suggestion, and capacity estimate.
报告撰稿人；2 种 mode（single-shot / chapter pipeline）+ 3 个 duty（chapter / dimension-outline / mission-outline / single-shot）
跨维度综合分析师；产出 insights / contradictions / gaps，喂给 Writer 落到报告
Mission 唯一最终负责人；M0 plan / M1 assess-research / M6 foreword / M7 sign-off 4 个 milestone 全程在场
跨维度对账专员；整合 findings、抽取事实、识别冲突、列出空白
单维度数据采集者；并发执行，专注证据驱动的结构化 finding 产出
主观质量评审员；3 种粒度（mission-review / mission-critic / dimension-quality）打分 + 给 critique
