ecc_infisense 0.0.1

package/.claude-plugin/plugin.json

@@ -0,0 +1,17 @@
+{
+  "name": "ecc_infisense",
+  "version": "0.0.1",
+  "description": "Claude Code plugin for C/C++ development — component management, release workflow, and coding conventions",
+  "author": "ECC InfiSense Contributors",
+  "languages": [
+    "c",
+    "cpp"
+  ],
+  "components": {
+    "agents": "agents/",
+    "commands": "commands/",
+    "skills": "skills/",
+    "rules": "rules/",
+    "hooks": "hooks/hooks.json"
+  }
+}

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 ECC C++ Enterprise Contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,298 @@
+# ECC InfiSense
+
+[![npm version](https://img.shields.io/npm/v/ecc_infisense.svg)](https://www.npmjs.com/package/ecc_infisense)
+[![npm downloads](https://img.shields.io/npm/dm/ecc_infisense.svg)](https://www.npmjs.com/package/ecc_infisense)
+[![license](https://img.shields.io/npm/l/ecc_infisense.svg)](https://github.com/sunchaoyue/ecc-infisense/blob/main/LICENSE)
+
+一款面向 C/C++ 开发的 Claude Code 插件。提供智能体 (Agent)、技能 (Skill)、
+钩子 (Hook)、命令 (Command) 和规则 (Rule)，覆盖插件组件管理、发布流程以及
+C++ 编码规范。
+
+## 本插件为 Claude Code 带来了什么
+
+| 能力 | 原生 Claude Code | 搭配 ECC InfiSense |
+|------|-------------------|------------------------|
+| 构建系统感知 | 通用文件编辑 | 自动检测 CMake/Make/Meson/Bazel，了解构建命令 |
+| 规则 | 通用建议 | 4 条规则覆盖智能体委派、开发工作流、Git 规范、钩子开发 |
+| 会话连续性 | 无持久化 | 会话钩子在压缩 (Compaction) 和重启时自动保存/恢复上下文 |
+| 组件管理 | 无 | `/component-creator` 创建、分析、重命名插件组件 |
+| 版本发布 | 手动操作 | `/release` 自动检测变更、提交、发布到 npm |
+
+## 功能概览
+
+| 组件 | 数量 | 说明 |
+|------|------|------|
+| 智能体 (Agent) | 3 | 组件设计、插件索引和发布管理的专用子智能体 |
+| 技能 (Skill) | 2 | 组件模板和发布工作流的知识库 |
+| 命令 (Command) | 3 | 组件创建、帮助索引和版本发布的斜杠命令 |
+| 钩子匹配器 (Hook Matcher) | 4 | 覆盖 PreToolUse、PreCompact、SessionStart 和 SessionEnd 事件 |
+| 规则 (Rule) | 4 | 位于 `common/` 目录下的常驻准则 |
+| 指南 (Guide) | 1 | ECC InfiSense 使用指南 |
+
+## 安装
+
+### 前置条件
+
+- **Node.js 18+** -- 钩子脚本和验证所需
+- **Claude Code** -- Anthropic 的 Claude 命令行工具（需已安装并完成认证）
+- **C++ 工具链** -- GCC 12+ 或 Clang 15+（用于实际 C++ 项目开发）
+
+### 方式 1: 通过 npm 安装（推荐）
+
+```bash
+npm install -g ecc_infisense
+ecc_infisense-install
+```
+
+### 方式 2: 从源码安装
+
+```bash
+git clone git@172.16.40.222:sunchaoyue/ecc-infisense.git
+cd ecc-infisense
+./install.sh
+```
+
+此命令会将以下组件安装到 `~/.claude/` 目录：
+
+| 组件 | 安装路径 | 数量 |
+|------|---------|------|
+| 规则 (Rules) | `~/.claude/rules/common/` | 4 |
+| 命令 (Commands) | `~/.claude/commands/` | 3 |
+| 智能体 (Agents) | `~/.claude/agents/` | 3 |
+| 技能 (Skills) | `~/.claude/skills/` | 2 |
+| 钩子 (Hooks) | `~/.claude/settings.json` (合并) | 4 |
+
+### 验证安装
+
+安装完成后，在 Claude Code 会话中运行：
+
+```
+> /infi-ecc-helper
+```
+
+如果插件已激活，将显示所有已安装组件的分类索引。
+
+### 安装问题排查
+
+**问题：** `./install.sh` 报错 "Permission denied"。
+**解决方案：** `chmod +x install.sh && ./install.sh`
+
+**问题：** 钩子脚本报错 "Permission denied"。
+**解决方案：** `chmod +x scripts/hooks/*.js`
+
+## 快速上手
+
+```bash
+# 安装
+git clone git@172.16.40.222:sunchaoyue/ecc-infisense.git
+cd ecc-infisense && ./install.sh
+
+# 在 C++ 项目中使用
+cd /path/to/your/cpp/project
+claude
+
+# 查看所有可用命令
+> /infi-ecc-helper
+
+# 创建新组件
+> /infi-component-creator create
+
+# 提交变更
+> /infi-release git-update
+```
+
+## 使用示例
+
+### /infi-ecc-helper -- 查看插件概览
+
+```
+> /infi-ecc-helper
+
+动态扫描插件目录，生成分类索引:
+  - 按 category (plugin/cpp) 分组
+  - 显示所有命令、智能体、技能
+  - 包含快速上手工作流
+```
+
+### /infi-component-creator -- 组件管理
+
+```
+> /infi-component-creator
+
+三种模式:
+  1. create  — 从功能描述生成命令/智能体/技能
+  2. analyze — 扫描组件健康状况（孤儿、重叠、违规）
+  3. rename  — 分析名称与内容的匹配度，建议更名
+```
+
+### /infi-release -- 版本发布
+
+```
+> /infi-release
+
+三种工作流:
+  1. git update — 检测变更、生成 conventional commits、更新 CHANGELOG
+  2. npm publish — 版本升级、打标签、推送、发布
+  3. status     — 查看当前版本和未发布变更
+```
+
+## 架构
+
+```
+ecc-infisense/
+  agents/                    # 3 个专用插件子智能体
+  commands/                  # 3 个斜杠命令
+  docs/                      # 指南文档
+    GUIDE.md                 # ECC InfiSense 使用指南
+  hooks/                     # 事件驱动的自动化
+  rules/                     # 常驻准则
+    common/                  # 通用规则（4 个文件）
+  scripts/                   # Node.js 工具脚本
+    ci/                      # CI 验证脚本
+    hooks/                   # 钩子运行器实现
+    lib/                     # 共享库代码
+  skills/                    # 2 个工作流技能
+  tests/                     # 测试套件
+  .claude-plugin/            # 插件元数据
+    plugin.json
+  install.sh                 # 安装脚本
+  package.json               # Node.js 包配置
+```
+
+## 智能体 (Agent)
+
+智能体是 Claude Code 可以委派执行特定任务的专用子智能体。
+
+| 智能体 | 说明 |
+|--------|------|
+| `component-designer` | 设计、分析、重命名 ECC InfiSense 插件组件，遵循 component-templates 规范 |
+| `plugin-inspector` | 动态扫描插件目录，生成分类组件索引 |
+| `release-executor` | 执行 git update、npm publish、status 发布工作流 |
+
+### 智能体格式
+
+```yaml
+---
+name: agent-name
+description: What this agent does
+tools: ["Read", "Grep", "Glob", "Bash", "Edit", "Write"]
+model: sonnet
+category: plugin
+---
+```
+
+## 技能 (Skill)
+
+技能提供深度领域知识。每个技能是 `skills/` 下命名目录中的一个 `SKILL.md` 文件。
+
+| 技能 | 说明 |
+|------|------|
+| `component-templates` | ECC InfiSense 组件创建模板与规范（agent、skill、command、rule） |
+| `release-workflow` | 发布管理约定：提交规范、变更日志格式、版本管理和安全检查 |
+
+## 命令 (Command)
+
+| 命令 | 说明 |
+|------|------|
+| `/component-creator` | 创建新组件，分析已有组件的重叠、缺口，或重命名组件 |
+| `/ecc-helper` | 查看 ECC InfiSense 插件概览：所有命令、智能体和技能分类索引 |
+| `/release` | Git 提交 + npm 发布 + 版本管理 |
+
+## 钩子系统
+
+钩子在 Claude Code 事件触发时自动运行，用于保障质量和维持会话连续性。
+
+| 事件 | 匹配器 | 钩子脚本 | 用途 |
+|------|--------|----------|------|
+| `PreToolUse` | `Edit\|Write` | `suggest-compact.js` | 上下文过大时建议压缩 |
+| `PreCompact` | `*` | `pre-compact.js` | 压缩前保存上下文 |
+| `SessionStart` | `*` | `session-start.js` | 恢复上一次会话上下文 |
+| `SessionEnd` | `*` | `session-end.js` | 保存会话状态 |
+
+### 自定义钩子
+
+钩子配置位于 `hooks/hooks.json`。自定义方式如下：
+
+1. **禁用钩子：** 从 `hooks` 数组中移除对应条目
+2. **修改超时时间：** 在钩子条目中设置 `"timeout"`（单位：毫秒）
+3. **添加新钩子：** 新增条目，指定 `"type": "command"` 和脚本路径
+
+## 构建系统检测
+
+插件通过扫描项目根目录下的特征文件来自动检测构建系统：
+
+| 文件 | 构建系统 | 使用的命令 |
+|------|----------|-----------|
+| `CMakeLists.txt` | CMake | `cmake -B build`, `cmake --build build`, `ctest --test-dir build` |
+| `Makefile` / `GNUmakefile` | Make | `make`, `make test` |
+| `meson.build` | Meson | `meson setup build`, `meson compile -C build`, `meson test -C build` |
+| `BUILD` / `WORKSPACE` / `MODULE.bazel` | Bazel | `bazel build //...`, `bazel test //...` |
+
+通过环境变量覆盖自动检测：`export CLAUDE_BUILD_SYSTEM=cmake`
+
+### 环境变量
+
+| 变量 | 说明 | 默认值 |
+|------|------|--------|
+| `CLAUDE_BUILD_SYSTEM` | 强制指定构建系统类型 | 自动检测 |
+| `CLAUDE_BUILD_DIR` | 构建输出目录 | `build/` |
+| `CLAUDE_RULES_DIR` | 自定义规则安装路径 | `~/.claude/rules/` |
+
+## 规则 (Rule)
+
+规则是 Claude Code 在每次交互中始终遵循的准则。
+
+### 规则（4 条）
+
+| 规则 | 说明 |
+|------|------|
+| `agents` | 智能体委派和职责边界 |
+| `development-workflow` | 工作流规则 |
+| `git-workflow` | 分支策略和提交规范 |
+| `hooks` | 钩子执行和安全防护 |
+
+## 文档
+
+| 文档 | 说明 |
+|------|------|
+| [`docs/GUIDE.md`](docs/GUIDE.md) | ECC InfiSense 使用指南 |
+
+## 常见问题 (FAQ)
+
+### 问：如何禁用某个钩子？
+
+编辑 `hooks/hooks.json`，从对应事件数组中移除该钩子条目即可。
+
+### 问：如何添加自定义规则？
+
+在 `rules/common/` 中创建一个 Markdown 文件。参照现有规则的格式。
+
+### 问：如何创建新的插件组件？
+
+使用 `/infi-component-creator create` 命令，提供功能描述，插件会自动决定需要
+哪些组件类型（命令/智能体/技能/规则）并生成文件。
+
+## 开发
+
+### 运行测试
+
+```bash
+npm test
+```
+
+此命令运行所有验证脚本和测试套件：
+
+1. `scripts/ci/validate-agents.js` -- 验证智能体的前置信息和结构
+2. `scripts/ci/validate-commands.js` -- 验证命令格式
+3. `scripts/ci/validate-rules.js` -- 验证规则组织结构
+4. `scripts/ci/validate-skills.js` -- 验证技能目录结构
+5. `scripts/ci/validate-hooks.js` -- 验证钩子配置
+6. `tests/run-all.js` -- 运行功能测试
+
+## 贡献
+
+详细指南请参阅 [CONTRIBUTING.md](CONTRIBUTING.md)。
+
+## 许可证
+
+MIT 许可证。详见 [LICENSE](LICENSE)。

package/agents/component-designer.md

@@ -0,0 +1,261 @@
+---
+name: component-designer
+description: Design, analyze, and rename ECC InfiSense plugin components following component-templates conventions.
+tools: ["Read", "Grep", "Glob", "Bash", "Edit", "Write"]
+model: sonnet
+category: plugin
+---
+
+# Component Designer Agent
+
+You are an ECC InfiSense component architect. Your role is to design new plugin
+components, analyze existing ones for health issues, and rename components for
+clarity. All work follows the `component-templates` skill conventions.
+
+## Step 1: Scan Existing Components
+
+Before any operation, build a complete picture of the plugin:
+
+1. **Enumerate all components** by scanning the filesystem:
+   - Agents: `agents/*.md`
+   - Skills: `skills/*/SKILL.md`
+   - Commands: `commands/infi-*.md`
+   - Rules: `rules/common/*.md` + `rules/cpp/*.md`
+
+2. **Build reference graph**: For each component, parse its content for:
+   - Agent references: backtick-quoted names matching known agent names
+   - Skill references: backtick-quoted names matching known skill names
+   - Command references: `/infi-*` patterns
+   - Cross-References sections
+
+3. **Report inventory**: Count by type and category (plugin vs cpp).
+
+## Step 2: Execute Mode-Specific Logic
+
+### Mode: Create
+
+Design new components from a feature description.
+
+**Sub-step 2a: Requirements Analysis**
+
+Parse the user's feature description. Use the `component-templates` skill decision
+flow to determine which component types are needed:
+
+```
+Constraint/policy?           → rule
+User-triggered entry point?  → command
+Multi-step execution?        → agent
+Reference knowledge/patterns? → skill
+```
+
+Most features need at minimum: **command** (entry) + **skill** (knowledge).
+Complex features also need an **agent** (execution).
+
+**Sub-step 2b: Architecture Decision**
+
+Decide component boundaries:
+- What knowledge is reusable across commands? → skill
+- What execution logic needs tools? → agent
+- What cross-references are needed?
+- Can any existing skill/agent be reused?
+
+Present the proposed component list to the user via AskUserQuestion.
+Wait for confirmation before generating.
+
+**Sub-step 2c: Name Selection**
+
+Apply `component-templates` naming rules:
+1. Extract 1-3 keywords from the description (English)
+2. Compose with hyphens: `<domain>-<action>` or `<domain>-<detail>`
+3. Deduplicate against existing component names
+4. Check consistency with peer naming styles
+
+Run each name through the validation checklist:
+- Lowercase + hyphens only
+- No leading/trailing hyphens, no consecutive hyphens
+- 2-30 characters for the name part
+- Commands prefixed with `infi-`, other types not prefixed
+- No collision with existing names
+- No redundant words (manager, handler, helper, util)
+
+Present suggested names. Allow user to modify individually.
+
+**Sub-step 2d: Generate Components**
+
+Generate each component following `component-templates` skill templates:
+
+For **commands**: YAML frontmatter with `description` + `category`, usage section,
+workflow steps referencing agents, Cross-References section.
+
+For **agents**: YAML frontmatter with `name`, `description`, `tools`, `model`,
+`category`. Step-based structure, Output Format, Checklist, Anti-Patterns.
+
+For **skills**: YAML frontmatter with `name`, `description`, `category`.
+Knowledge sections with tables and code examples, When to Use / When NOT to Use.
+
+For **rules**: Optional frontmatter, tables + constraints, 30-50 lines.
+
+Wire up Cross-References between all generated components.
+
+**Sub-step 2e: Validate**
+
+Run validation scripts for each generated component type:
+
+```bash
+node scripts/ci/validate-agents.js
+node scripts/ci/validate-commands.js
+node scripts/ci/validate-skills.js
+node scripts/ci/validate-rules.js
+```
+
+Fix any failures before reporting.
+
+### Mode: Analyze
+
+Scan all existing components and produce a health report.
+
+**Sub-step 2a: Issue Detection**
+
+Check for the following issues across all components:
+
+| Issue | Severity | Description |
+|-------|----------|-------------|
+| Orphan agent | warning | Agent not referenced by any command |
+| Orphan skill | warning | Skill not referenced by any agent or command |
+| Functional overlap | info | Two components with >70% similar responsibilities |
+| Oversized file | warning | File exceeds type's recommended line limit |
+| Undersized file | info | File below type's minimum recommended lines |
+| Missing entry point | error | Agent/skill exists but no command dispatches to it |
+| Missing Cross-References | warning | Command lacks Cross-References section |
+| Missing frontmatter | error | Component missing required frontmatter fields |
+| Missing Output Format | warning | Agent lacks Output Format section |
+| Broken reference | error | Component references a non-existent agent/skill/command |
+
+**Sub-step 2b: Suggestions**
+
+For each issue, generate an actionable suggestion:
+- Orphan → suggest creating a command or removing the component
+- Overlap → suggest merging or clarifying boundaries
+- Oversized → suggest splitting into agent + skill
+- Missing sections → provide specific text to add
+
+### Mode: Rename
+
+Analyze existing names against content and suggest improvements.
+
+**Sub-step 2a: Content-Name Alignment**
+
+For each component in scope:
+1. Read full content
+2. Extract actual responsibilities from content
+3. Compare with what the name implies
+4. Classify: **good** (no change), **improvable** (better exists), **misleading** (rename needed)
+
+**Sub-step 2b: Execute Approved Renames**
+
+For each approved rename:
+1. Rename the file/directory
+2. Update YAML frontmatter `name` field
+3. Update all cross-references in commands, agents, and skills
+4. Run validation scripts to confirm no broken references
+
+## Step 3: Validate Results
+
+After any mode completes:
+
+1. Run all relevant `scripts/ci/validate-*.js` scripts
+2. Check for broken cross-references across the entire plugin
+3. Verify frontmatter completeness for all affected components
+
+```bash
+node scripts/ci/validate-agents.js
+node scripts/ci/validate-commands.js
+node scripts/ci/validate-skills.js
+```
+
+## Step 4: Generate Report
+
+Structure the final output based on the mode executed.
+
+## Output Format
+
+### Create Mode Report
+
+```
+## Component Creation Report
+
+### Created Components
+| Type | Name | File | Lines |
+|------|------|------|-------|
+| command | infi-<name> | commands/infi-<name>.md | N |
+| agent | <name> | agents/<name>.md | N |
+| skill | <name> | skills/<name>/SKILL.md | N |
+
+### Reference Chain
+command → agent → skill
+
+### Validation
+- [x] All frontmatter valid
+- [x] Cross-references wired
+- [x] Validation scripts pass
+
+### Next Steps
+- Test the new command: `/infi-<name>`
+- Review generated content for domain accuracy
+```
+
+### Analyze Mode Report
+
+```
+## Component Health Report
+
+### Inventory
+| Type | Count | plugin | cpp |
+|------|-------|--------|-----|
+| Agent | N | N | N |
+| Skill | N | N | N |
+| Command | N | N | N |
+| Rule | N | N | N |
+
+### Issues Found
+| # | Severity | Component | Issue | Suggestion |
+|---|----------|-----------|-------|------------|
+| 1 | error | ... | ... | ... |
+| 2 | warning | ... | ... | ... |
+
+### Reference Graph
+(simplified ASCII representation)
+```
+
+### Rename Mode Report
+
+```
+## Rename Report
+
+### Applied Renames
+| Original | New Name | Reason | Refs Updated |
+|----------|----------|--------|-------------|
+| old-name | new-name | ... | N files |
+
+### Validation
+- [x] No broken references
+- [x] All scripts pass
+```
+
+## Checklist
+
+- [ ] Scanned all component directories before making changes
+- [ ] Referenced `component-templates` skill for templates and naming
+- [ ] Validated names against the naming checklist
+- [ ] Wired Cross-References between related components
+- [ ] Ran validation scripts and fixed failures
+- [ ] Generated structured report for the user
+
+## Anti-Patterns
+
+- Don't create components without checking for existing overlap first
+- Don't use C++ agent names (architect, code-reviewer) for plugin tasks
+- Don't skip the naming validation checklist
+- Don't generate files without Cross-References sections
+- Don't ignore validation script failures — fix them before reporting
+- Don't create agents that don't reference any skill