feat: add nabledge-1.x knowledge generation infrastructure (#189)

.claude/skills/nabledge-1.2/SKILL.md

@@ -0,0 +1,74 @@
+---
+name: nabledge-1.2
+description: Provides Nablarch 1.2 framework knowledge and code analysis capabilities. Use when developing Nablarch applications, implementing features, reviewing code, or answering questions about Nablarch 1.2.
+user-invocable: false
+disable-model-invocation: true
+---
+
+# Nabledge-1.2: Nablarch 1.2 Knowledge Base
+
+Knowledge base and code analysis tool for Nablarch 1.2 framework.
+
+## Usage
+
+**Interactive mode**:
+```
+nabledge-1.2
+```
+
+**Direct knowledge search**:
+```
+nabledge-1.2 "<question>"
+```
+
+**Direct code analysis**:
+```
+nabledge-1.2 code-analysis
+```
+
+## Execution Instructions
+
+### Step 0: Determine Workflow
+
+**No arguments** (`nabledge-1.2`):
+- Show greeting
+- Ask user to choose: Knowledge Search or Code Analysis
+
+**Text argument** (`nabledge-1.2 "<question>"`):
+- Execute `workflows/qa.md` to answer question
+- This workflow orchestrates _knowledge-search pipeline
+
+**"code-analysis" argument** (`nabledge-1.2 code-analysis`):
+- Execute `workflows/code-analysis.md` to analyze user's code
+
+## Critical Constraints
+
+**Knowledge files only**: Answer using ONLY information in `knowledge/*.json`. DO NOT use external knowledge or LLM training data.
+
+**When knowledge is missing**:
+- Output: "この情報は知識ファイルに含まれていません"
+- List related available knowledge from `knowledge/index.toon`
+- DO NOT answer from external knowledge
+
+## Error Handling
+
+**Knowledge not found** (Knowledge Search):
+- Message: "この情報は知識ファイルに含まれていません"
+- List related entries from index.toon
+
+**Target code not found** (Code Analysis):
+- Message: "指定されたコードが見つかりませんでした"
+- Show search patterns used
+- Ask for clarification
+
+**Workflow execution failure**:
+- Inform which step failed
+- Show error details
+- Suggest retry or alternative
+
+## Knowledge Structure
+
+**Files**: `knowledge/features/`, `knowledge/checks/`, `knowledge/releases/`
+**Index**: `knowledge/index.toon` (all entries with search hints)
+**Schemas**: `schemas/*.json` (JSON validation schemas)
+**Scripts**: `scripts/*.sh` (pre-built processing scripts)

.claude/skills/nabledge-1.2/plugin/CHANGELOG.md

@@ -0,0 +1,7 @@
+# 変更履歴
+
+nabledge-1.2プラグインの主な変更内容を記録しています。
+
+フォーマットは [Keep a Changelog](https://keepachangelog.com/en/1.0.0/) に基づいています。
+
+## [Unreleased]

.claude/skills/nabledge-1.2/plugin/plugin.json

@@ -0,0 +1,13 @@
+{
+  "name": "nabledge-1.2",
+  "version": "0.1",
+  "description": "Nablarch 1.2 skill for AI-assisted development",
+  "author": {
+    "name": "Nablarch"
+  },
+  "license": "Apache-2.0",
+  "repository": "https://github.com/nablarch/nabledge",
+  "keywords": [
+    "nablarch"
+  ]
+}

.claude/skills/nabledge-1.3/SKILL.md

@@ -0,0 +1,74 @@
+---
+name: nabledge-1.3
+description: Provides Nablarch 1.3 framework knowledge and code analysis capabilities. Use when developing Nablarch applications, implementing features, reviewing code, or answering questions about Nablarch 1.3.
+user-invocable: false
+disable-model-invocation: true
+---
+
+# Nabledge-1.3: Nablarch 1.3 Knowledge Base
+
+Knowledge base and code analysis tool for Nablarch 1.3 framework.
+
+## Usage
+
+**Interactive mode**:
+```
+nabledge-1.3
+```
+
+**Direct knowledge search**:
+```
+nabledge-1.3 "<question>"
+```
+
+**Direct code analysis**:
+```
+nabledge-1.3 code-analysis
+```
+
+## Execution Instructions
+
+### Step 0: Determine Workflow
+
+**No arguments** (`nabledge-1.3`):
+- Show greeting
+- Ask user to choose: Knowledge Search or Code Analysis
+
+**Text argument** (`nabledge-1.3 "<question>"`):
+- Execute `workflows/qa.md` to answer question
+- This workflow orchestrates _knowledge-search pipeline
+
+**"code-analysis" argument** (`nabledge-1.3 code-analysis`):
+- Execute `workflows/code-analysis.md` to analyze user's code
+
+## Critical Constraints
+
+**Knowledge files only**: Answer using ONLY information in `knowledge/*.json`. DO NOT use external knowledge or LLM training data.
+
+**When knowledge is missing**:
+- Output: "この情報は知識ファイルに含まれていません"
+- List related available knowledge from `knowledge/index.toon`
+- DO NOT answer from external knowledge
+
+## Error Handling
+
+**Knowledge not found** (Knowledge Search):
+- Message: "この情報は知識ファイルに含まれていません"
+- List related entries from index.toon
+
+**Target code not found** (Code Analysis):
+- Message: "指定されたコードが見つかりませんでした"
+- Show search patterns used
+- Ask for clarification
+
+**Workflow execution failure**:
+- Inform which step failed
+- Show error details
+- Suggest retry or alternative
+
+## Knowledge Structure
+
+**Files**: `knowledge/features/`, `knowledge/checks/`, `knowledge/releases/`
+**Index**: `knowledge/index.toon` (all entries with search hints)
+**Schemas**: `schemas/*.json` (JSON validation schemas)
+**Scripts**: `scripts/*.sh` (pre-built processing scripts)

.claude/skills/nabledge-1.3/plugin/CHANGELOG.md

@@ -0,0 +1,7 @@
+# 変更履歴
+
+nabledge-1.3プラグインの主な変更内容を記録しています。
+
+フォーマットは [Keep a Changelog](https://keepachangelog.com/en/1.0.0/) に基づいています。
+
+## [Unreleased]

.claude/skills/nabledge-1.3/plugin/plugin.json

@@ -0,0 +1,13 @@
+{
+  "name": "nabledge-1.3",
+  "version": "0.1",
+  "description": "Nablarch 1.3 skill for AI-assisted development",
+  "author": {
+    "name": "Nablarch"
+  },
+  "license": "Apache-2.0",
+  "repository": "https://github.com/nablarch/nabledge",
+  "keywords": [
+    "nablarch"
+  ]
+}

.claude/skills/nabledge-1.4/SKILL.md

@@ -0,0 +1,74 @@
+---
+name: nabledge-1.4
+description: Provides Nablarch 1.4 framework knowledge and code analysis capabilities. Use when developing Nablarch applications, implementing features, reviewing code, or answering questions about Nablarch 1.4.
+user-invocable: false
+disable-model-invocation: true
+---
+
+# Nabledge-1.4: Nablarch 1.4 Knowledge Base
+
+Knowledge base and code analysis tool for Nablarch 1.4 framework.
+
+## Usage
+
+**Interactive mode**:
+```
+nabledge-1.4
+```
+
+**Direct knowledge search**:
+```
+nabledge-1.4 "<question>"
+```
+
+**Direct code analysis**:
+```
+nabledge-1.4 code-analysis
+```
+
+## Execution Instructions
+
+### Step 0: Determine Workflow
+
+**No arguments** (`nabledge-1.4`):
+- Show greeting
+- Ask user to choose: Knowledge Search or Code Analysis
+
+**Text argument** (`nabledge-1.4 "<question>"`):
+- Execute `workflows/qa.md` to answer question
+- This workflow orchestrates _knowledge-search pipeline
+
+**"code-analysis" argument** (`nabledge-1.4 code-analysis`):
+- Execute `workflows/code-analysis.md` to analyze user's code
+
+## Critical Constraints
+
+**Knowledge files only**: Answer using ONLY information in `knowledge/*.json`. DO NOT use external knowledge or LLM training data.
+
+**When knowledge is missing**:
+- Output: "この情報は知識ファイルに含まれていません"
+- List related available knowledge from `knowledge/index.toon`
+- DO NOT answer from external knowledge
+
+## Error Handling
+
+**Knowledge not found** (Knowledge Search):
+- Message: "この情報は知識ファイルに含まれていません"
+- List related entries from index.toon
+
+**Target code not found** (Code Analysis):
+- Message: "指定されたコードが見つかりませんでした"
+- Show search patterns used
+- Ask for clarification
+
+**Workflow execution failure**:
+- Inform which step failed
+- Show error details
+- Suggest retry or alternative
+
+## Knowledge Structure
+
+**Files**: `knowledge/features/`, `knowledge/checks/`, `knowledge/releases/`
+**Index**: `knowledge/index.toon` (all entries with search hints)
+**Schemas**: `schemas/*.json` (JSON validation schemas)
+**Scripts**: `scripts/*.sh` (pre-built processing scripts)

.claude/skills/nabledge-1.4/docs/README.md

@@ -0,0 +1,3 @@
+# Nablarch 6 ドキュメント
+
+0 ページ

.claude/skills/nabledge-1.4/knowledge/index.toon

@@ -0,0 +1,3 @@
+# Nabledge-1.4 Knowledge Index
+
+files[0,]{title,type,category,processing_patterns,path}:

.claude/skills/nabledge-1.4/plugin/CHANGELOG.md

@@ -0,0 +1,7 @@
+# 変更履歴
+
+nabledge-1.4プラグインの主な変更内容を記録しています。
+
+フォーマットは [Keep a Changelog](https://keepachangelog.com/en/1.0.0/) に基づいています。
+
+## [Unreleased]

.claude/skills/nabledge-1.4/plugin/plugin.json

@@ -0,0 +1,13 @@
+{
+  "name": "nabledge-1.4",
+  "version": "0.1",
+  "description": "Nablarch 1.4 skill for AI-assisted development",
+  "author": {
+    "name": "Nablarch"
+  },
+  "license": "Apache-2.0",
+  "repository": "https://github.com/nablarch/nabledge",
+  "keywords": [
+    "nablarch"
+  ]
+}

.pr/00189/review-by-qa-engineer.md

@@ -0,0 +1,56 @@
+# Expert Review: QA Engineer
+
+**Date**: 2026-03-13
+**Reviewer**: AI Agent as QA Engineer
+**Files Reviewed**: 1 file
+
+## Overall Assessment
+
+**Rating**: 4/5
+**Summary**: `TestV1xVersions` is well-structured and consistent with existing test patterns. Tests are clear, focused, and exercise correct routing logic for all three v1.x versions.
+
+## Key Issues
+
+### Medium Priority
+
+1. **Missing `--resume` coverage for v1.x gen**
+   - Description: `TestGenCommand` has `test_gen_resume_skips_clean` for v6 but no equivalent for v1.x.
+   - Suggestion: Add parametrized `test_gen_v1x_resume_skips_clean` asserting 1 command with `scripts/run.py`.
+   - Decision: **Implement Now** — added `test_gen_v1x_resume_skips_clean` test.
+   - Reasoning: Closes a real behavioral gap with minimal effort.
+
+2. **regen/fix tests don't assert `scripts/run.py`**
+   - Description: Assertions check flags but not which script is called, unlike gen test.
+   - Suggestion: Add `assert "scripts/run.py" in lines[0]` to regen and fix tests.
+   - Decision: **Implement Now** — added to both tests.
+   - Reasoning: Tightens routing verification consistently with gen test.
+
+### Low Priority
+
+3. **Split dual assertion in test_gen_v1x_calls_clean_then_run**
+   - Description: Single `and`-joined assertion makes failure messages less specific.
+   - Suggestion: Split into two separate assertions.
+   - Decision: **Defer** — correct behavior, style preference only.
+   - Reasoning: Low value change; existing test pattern uses same style.
+
+4. **No `returncode` assertion in success cases**
+   - Description: Tests pass even if kc.sh exits with error after printing CMD lines.
+   - Suggestion: Add `assert result.returncode == 0` — ideally backported to whole file.
+   - Decision: **Defer** — whole-file refactor; follow-up.
+   - Reasoning: Existing tests have same gap; out of scope for this PR.
+
+## Positive Aspects
+
+- Parametrize approach efficiently covers all three versions without duplication.
+- Version argument assertion in regen/fix is an improvement over existing v6 tests.
+- New class is self-contained and doesn't interfere with existing classes.
+- Test naming follows established `test_{command}_{context}_{behavior}` convention.
+- Failure message in gen test includes version for clear parametrized failure diagnosis.
+
+## Recommendations
+
+- Add `returncode` assertions as a follow-up for the whole file.
+
+## Files Reviewed
+
+- `tools/knowledge-creator/tests/ut/test_kc_sh.py` (tests)