current-state-analysis · diff
git:20260812.cc20b07 to git:20260812.92c38df
104 added, 82 removed. Audit A to A.
---
name: current-state-analysis
description: >-
- Run current-state analysis and problem-identification frameworks (現状分析・
- 課題の特定) on user-supplied material and render the results as slides from
- the analysis template pack. Covers PEST, Five Forces, business-process
- pain-point mapping, logic tree, KPI tree, why-why (5 Whys), fishbone,
- Pareto, As-Is/To-Be gap analysis and an impact-effort priority matrix, plus
- SWOT / 3C from the marketing-analysis pack. Use when asked to 現状分析,
- 課題を特定・構造化, 真因分析, As-Is/To-Be 整理, 課題の優先順位付け, or to
- turn hearing notes / data into analysis slides. Route new-template authoring
- to analysis-template-creator, deck generation mechanics to google-slides,
- and visual QA to slide-qa.
+ Run current-state analysis and problem-identification frameworks
+ (current-state analysis, problem identification) on user-supplied material
+ and render the results as slides from the analysis template pack. Covers
+ PEST, Five Forces, business-process pain-point mapping, logic tree, KPI
+ tree, why-why (5 Whys), fishbone, Pareto, As-Is/To-Be gap analysis and an
+ impact-effort priority matrix, plus SWOT / 3C from the marketing-analysis
+ pack. Use when asked to 現状分析, 課題を特定・構造化, 真因分析, As-Is/To-Be
+ 整理, 課題の優先順位付け, or to turn hearing notes / data into analysis
+ slides. Route new-template authoring to analysis-template-creator, deck
+ generation mechanics to google-slides, and visual QA to slide-qa.
---
+ *[日本語](SKILL.ja.md)*
- # Current-State Analysis（現状分析・課題の特定）
+ # Current-State Analysis (Current-State Analysis & Problem Identification)
- ユーザーが持ち込む材料（ヒアリングメモ・業務資料・データ）に分析フレーム
- ワークを適用し、その結果を `slide-templates/analysis` パック（+ 一部
- `marketing-analysis` パック）のスライドに落とす。
+ Applies an analysis framework to material the user brings in (hearing notes,
+ business documents, data) and renders the result as slides from the
+ `slide-templates/analysis` pack (plus, in some cases, the
+ `marketing-analysis` pack).
- 作業ディレクトリは slide-forge ルート。コマンドは `.venv/bin/python`。
+ The working directory is the slide-forge root. The command is
+ `.venv/bin/python`.
## Boundaries
- | 依頼 | 行き先 |
+ | Request | Route |
|---|---|
- | 現状分析・課題特定を実行し、スライドにする | このスキル |
- | 分析フレームワークのテンプレートを新設・変更する | `analysis-template-creator` |
- | SWOT / 3C / ポジショニングマップ等の市場分析 1 枚だけ | `google-slides-template` + marketing-analysis パック |
- | デッキ生成そのものの仕組み・マスター選択 | `google-slides` / `google-slides-template` |
- | 生成したデッキの目視検査 | `slide-qa` |
- | 商談・アカウント固有の課題整理（台帳あり） | `scalar-account-plan` 系 |
+ | Perform a current-state analysis / problem identification and turn it into slides | this skill |
+ | Create or change a template for an analysis framework | `analysis-template-creator` |
+ | A single market-analysis slide such as SWOT / 3C / positioning map | `google-slides-template` + the marketing-analysis pack |
+ | The mechanics of deck generation itself / master selection | `google-slides` / `google-slides-template` |
+ | Visual inspection of a generated deck | `slide-qa` |
+ | Deal- or account-specific problem organization (with an account ledger) | the `scalar-account-plan` family |
- ## 手法とテンプレートの対応
+ ## Mapping methods to templates
- 分析は「環境 → 業務 → 構造化 → 課題定義 → 優先順位」の順に深くなる。
- **全部やらない。** 問いに答えるのに必要な段だけ選ぶ。
+ Analysis deepens in the order "environment → business process → structuring
+ → problem definition → prioritization." **Do not do all of it.** Choose only
+ the stages needed to answer the question at hand.
- | 段 | 答える問い | 手法 | テンプレート |
+ | Stage | Question answered | Method | Template |
|---|---|---|---|
- | 1. 環境 | 外部要因は何か | PEST | `pest-analysis` |
- | 1. 環境 | 業界の競争圧力はどこから | 5フォース | `five-forces` |
- | 1. 環境 | 強み・弱み × 機会・脅威 | SWOT | `swot-analysis`（marketing-analysis） |
- | 1. 環境 | 顧客・競合・自社 | 3C | `three-c-analysis`（marketing-analysis） |
- | 2. 業務 | どの工程に問題が集中しているか | 業務フロー + ペイン | `process-painpoints` |
- | 3. 構造化 | 問題は何に分解できるか | ロジックツリー | `logic-tree` |
- | 3. 構造化 | 未達はどの指標のせいか | KPI ツリー | `kpi-tree` |
- | 3. 構造化 | 真因は何か（単線） | なぜなぜ分析 | `why-why` |
- | 3. 構造化 | 原因の仮説はどの系統か | 特性要因図 | `fishbone-diagram` |
- | 3. 構造化 | どの要因が大半を占めるか | パレート分析 | `pareto-analysis` |
- | 4. 課題定義 | 現状と理想の差は何か | ギャップ分析 | `gap-analysis` |
- | 5. 優先順位 | どれから着手するか | 優先順位マトリクス | `priority-matrix` |
+ | 1. Environment | What are the external factors? | PEST | `pest-analysis` |
+ | 1. Environment | Where does industry competitive pressure come from? | Five Forces | `five-forces` |
+ | 1. Environment | Strengths/weaknesses × opportunities/threats | SWOT | `swot-analysis` (marketing-analysis) |
+ | 1. Environment | Customer, competitor, company | 3C | `three-c-analysis` (marketing-analysis) |
+ | 2. Business process | Which process steps concentrate the problems? | Business flow + pain points | `process-painpoints` |
+ | 3. Structuring | What can the problem be broken down into? | Logic tree | `logic-tree` |
+ | 3. Structuring | Which metric is responsible for the shortfall? | KPI tree | `kpi-tree` |
+ | 3. Structuring | What is the root cause (single chain)? | Why-why analysis | `why-why` |
+ | 3. Structuring | Which family does the causal hypothesis belong to? | Fishbone diagram | `fishbone-diagram` |
+ | 3. Structuring | Which factor accounts for the majority? | Pareto analysis | `pareto-analysis` |
+ | 4. Problem definition | What is the gap between current and ideal state? | Gap analysis | `gap-analysis` |
+ | 5. Prioritization | Which should be tackled first? | Priority matrix | `priority-matrix` |
- 定番の組み合わせ（業務・IT 改革系）:
- `process-painpoints` →（`pareto-analysis` か `logic-tree`）→ `why-why` →
- `gap-analysis` → `priority-matrix`。戦略系は 1 段目（PEST / 5F / 3C / SWOT）
- から `gap-analysis` へ。
+ A standard combination (for business/IT transformation work):
+ `process-painpoints` → (`pareto-analysis` or `logic-tree`) → `why-why` →
+ `gap-analysis` → `priority-matrix`. For strategy-oriented work, go from
+ stage 1 (PEST / 5F / 3C / SWOT) straight to `gap-analysis`.
- 各テンプレートの入力スロットと制約は
- `slide-templates/analysis/<id>/template.json` が正で、
- `.venv/bin/python scripts/list_slide_templates.py --tag analysis` で一覧できる。
- このスキルでスロット定義を再掲しない。
+ The input slots and constraints for each template are authoritatively
+ defined in `slide-templates/analysis/<id>/template.json`, and can be listed
+ with `.venv/bin/python scripts/list_slide_templates.py --tag analysis`. This
+ skill does not restate the slot definitions.
## Workflow
### 1. Intake
- 一度にまとめて確認する:
+ Confirm all of the following in one pass:
- - 答えるべき問い（何を決めるための分析か。決裁・稟議・改善企画…）;
- - 材料（ヒアリングメモ・データ・公開資料）と、その**出典・鮮度**;
- - 分析の深さ（環境〜優先順位のどの段まで要るか）;
- - 出力形態（単スライド / 分析章まるごと / 既存デッキへの追加）。
+ - the question to be answered (what is this analysis for — a decision, an
+ approval, an improvement proposal…);
+ - the material (hearing notes, data, public documents) and its **source and
+ freshness**;
+ - the depth of analysis (which stage, from environment to prioritization, is
+ actually needed);
+ - the output form (a single slide / a whole analysis chapter / an addition
+ to an existing deck).
- 材料が無い分析は**引き受けない**。不足があれば、埋めるためのヒアリング
- 項目・調査項目のリストを先に返す（それ自体が成果物になる）。
+ **Do not take on** an analysis with no material. If something is missing,
+ first return a list of hearing items / research items needed to fill the
+ gap — that list is itself a deliverable.
- ### 2. 分析の実行
+ ### 2. Running the analysis
- フレームワークは思考の型であって、埋め草の型ではない:
+ A framework is a thinking template, not a filler template:
- - **事実と解釈を分ける。** 図の中身（工程・件数・環境要因・原因）は材料に
- ある事実だけ。解釈は `insight`（so_what）と `title` に置く。
- - 材料に無いことを推測で埋めるときは（仮説）と明示し、検証方法を添える。
- - 数値はすべて出典・期間・母数を `source` スロットに書く。書けない数値は
- 載せない。
- - 各テンプレートの `guardrails` は手法の誤用パターン（相関を因果と読む、
- 「人の不注意」で止める等）を列挙している。**分析中に読み、従う。**
+ - **Separate fact from interpretation.** The content of the diagram (process
+ steps, counts, environmental factors, causes) must be facts drawn from the
+ material only. Interpretation goes into `insight` (so_what) and `title`.
+ - When filling a gap not covered by the material with an inference, mark it
+ explicitly as (hypothesis) and attach a verification method.
+ - Every figure must state its source, period, and base in the `source` slot.
+ Do not include a figure that cannot be sourced this way.
+ - Each template's `guardrails` enumerate misuse patterns for the method
+ (reading correlation as causation, stopping at "human carelessness," etc).
+ **Read them during the analysis, and follow them.**
- ### 3. スロット JSON の作成と検証
+ ### 3. Authoring and validating the slot JSON
- 手法ごとに `example.json` と同じ形の入力 JSON を作り、1 枚ずつ検証する:
+ For each method, build an input JSON in the same shape as `example.json`,
+ and validate it one page at a time:
```bash
.venv/bin/python scripts/render_slide_template.py \
--template <id> --data <data.json> --out out/<n>_<id>.json
```
- 複数枚は `assemble_spec.py` で 1 つのデッキ仕様に束ね、オフラインで監査する:
+ For multiple pages, bundle them into a single deck spec with
+ `assemble_spec.py` and audit it offline:
```bash
- .venv/bin/python scripts/assemble_spec.py out/*_*.json --out out/analysis-deck.json --title <題名>
+ .venv/bin/python scripts/assemble_spec.py out/*_*.json --out out/analysis-deck.json --title <title>
.venv/bin/python scripts/build_deck.py --template templates/<master>.json \
--spec out/analysis-deck.json --dry-run --strict
```
- 監査指摘（はみ出し・重なり・折り返し）は**データ側を短く**して直す。
- テンプレートやプリミティブを変えたくなったら `analysis-template-creator` へ。
+ Fix audit findings (overflow, overlap, wrapping) by **shortening the data
+ side**. If you find yourself wanting to change the template or a primitive,
+ go to `analysis-template-creator`.
- ### 4. 生成と QA
+ ### 4. Generation and QA
- 生成の実行・保存先・タイトルの流儀は `google-slides` / `google-slides-template`
- に従う。生成後は `slide-qa` で目視検査し、QA サムネイルを後始末する。
+ Follow `google-slides` / `google-slides-template` for the conventions on how
+ to run generation, where to save, and titling. After generation, do a visual
+ inspection with `slide-qa` and clean up the QA thumbnails afterward.
### 5. Report
- - 使った手法と、その選定理由（どの問いに答えるためか）;
- - 各スライドの主張（title）と、根拠の出典;
- - 材料不足で（仮説）扱いにした箇所と、その検証方法;
- - デッキ URL と QA 結果（生成まで実行した場合）。
+ - The methods used and the rationale for choosing them (which question each
+ answers);
+ - Each slide's claim (title) and the source it is based on;
+ - Where material was insufficient and treated as (hypothesis), and how it
+ would be verified;
+ - The deck URL and QA result (if generation was carried through).
## Safety and quality rules
- - **分析結果を捏造しない。** インタビュー発言・件数・工数は実データのみ。
- デモ用は example.json の流儀に合わせ、source に「サンプル」と明記する。
- - 記述（descriptive）・診断（diagnostic）・因果（causal）の主張を混ぜない。
- パレート図は「集中している」までで、「これが原因」は why-why / 検証の仕事。
- - 顧客・個人が特定できる分析（体制・ペインの発言者等）は社内資料。顧客に
- 渡す版は `scalar-ae-materials` の露出チェックの考え方に従って別に作る。
- - 生成物・QA ファイルは `out/` 以下に置く（コミットしない）。
+ - **Never fabricate analysis results.** Interview statements, counts, and
+ effort figures must be real data only. For demo purposes, follow the
+ example.json convention and mark the source explicitly as "sample."
+ - Do not mix descriptive, diagnostic, and causal claims. A Pareto chart
+ supports "this is concentrated" — "this is the cause" is the job of
+ why-why / verification.
+ - An analysis that can identify a customer or an individual (organizational
+ structure, who said what about a pain point) is internal material. A
+ version handed to the customer must be produced separately, following the
+ exposure-check approach used in `scalar-ae-materials`.
+ - Keep generated artifacts and QA files under `out/` (do not commit them).