Add Ontoly Software Graph skill

README.md

@@ -17,8 +17,8 @@
 ## 📊 统计数据
 
 - **官方技能**: 182 个（来自 awesome-agent-skills，自动爬取）
-- **本地技能**: 61 个（25核心 + 30子技能 + 6系统内置）
-- **技能总数**: 243 个（官方 + 本地）
+- **本地技能**: 62 个（26核心 + 30子技能 + 6系统内置）
+- **技能总数**: 244 个（官方 + 本地）
 - **备份覆盖率**: 100%（71个压缩包，完整备份）
 - **自动更新**: 每24小时自动爬取最新技能
 
@@ -54,7 +54,7 @@
 - **Better Auth** (3个) - best-practices, commands, create-auth
 - **其他团队** - Tinybird, Remotion, Inngest 等
 
-## 💾 本地技能库（61个）
+## 💾 本地技能库（62个）
 
 ### 📝 内容创作与发布（10个）
 - **content-creation-publisher** ⭐⭐⭐⭐⭐ - 内容创作与发布全流程
@@ -112,6 +112,9 @@
 - **law-to-markdown** ⭐⭐ - 法律文档转换
 - **stock-analysis** ⭐⭐⭐ - 股票个股分析
 
+### 💻 软件开发与代码理解（1个）
+- **ontoly-software-graph** ⭐⭐⭐⭐ - 基于 Ontoly Software Graph 的架构审查、请求追踪、依赖分析与影响分析
+
 ### 🤝 智能体协作（3个）
 - **agent-team** ⭐⭐⭐ - 智能体团队协作
 - **multi-agent-meeting** ⭐⭐ - 多智能体会议

skills/ontoly-software-graph/README.md

@@ -0,0 +1,42 @@
+# Ontoly Software Graph
+
+使用 Ontoly 的确定性 Software Graph 和 MCP 能力，让 AI 编程代理在搜索源码之前，优先基于图谱证据理解代码库。
+
+## 适用场景
+
+- 架构审查
+- 请求链路追踪
+- 依赖分析
+- 重构影响分析
+- 配置与环境变量审计
+- 服务、模块、控制器、路由定位
+- 代码库 onboarding
+
+## 核心原则
+
+1. 先确认 Ontoly graph 是否存在。
+2. 如果缺失且允许本地分析，运行 `ontoly build .`。
+3. 优先使用 Ontoly CLI 或 MCP 查询 graph。
+4. 只有当 graph 无法回答、过期或不完整时，才回退到源码搜索。
+5. 回答必须包含节点、边、源码位置、诊断或框架分析等证据。
+
+## 安装 Ontoly
+
+```bash
+npm install -D ontoly
+```
+
+或使用 pnpm：
+
+```bash
+pnpm add -D ontoly
+```
+
+## 示例问题
+
+- Explain this repository.
+- Trace the login request.
+- Which service owns authentication?
+- What breaks if I remove `UserRepository`?
+- Which packages depend on `AuthModule`?
+

skills/ontoly-software-graph/SKILL.md

@@ -0,0 +1,73 @@
+---
+name: ontoly-software-graph
+description: Use Ontoly's deterministic Software Graph and MCP capabilities for architecture review, request tracing, dependency analysis, configuration lookup, and impact analysis before falling back to source-file search.
+---
+
+# Ontoly Software Graph
+
+Use this skill when a coding agent needs evidence-backed software understanding from an Ontoly graph before searching repository files directly.
+
+## When to Use
+
+- Explaining a repository architecture
+- Tracing a request, route, controller, service, or dependency path
+- Finding owners of services, modules, routes, configuration, or environment variables
+- Reviewing dependency impact before a refactor
+- Auditing dead code, cycles, unresolved imports, graph quality, or semantic coverage
+- Preparing documentation, onboarding notes, or architecture review from graph evidence
+
+## Required Workflow
+
+1. Check whether an Ontoly graph already exists by looking for `.ontoly/`, `SoftwareGraph.json`, `diagnostics.json`, validation reports, or an Ontoly MCP configuration.
+2. If no graph exists and the user permits local analysis, run:
+
+   ```bash
+   ontoly build .
+   ```
+
+3. Inspect graph health before answering: diagnostics, graph hash, semantic coverage, trust or quality score, framework detection, and generation timestamp.
+4. Prefer Ontoly CLI or MCP capabilities for graph questions instead of scanning source files first.
+5. Use repository search only when Ontoly cannot answer, the graph is stale, the graph is incomplete, or the user explicitly asks for source-level verification.
+6. Always cite graph evidence in the answer: node IDs, edge types, file paths, source locations, diagnostics, or framework analyzer output.
+7. State confidence from graph evidence. Do not guess confidence.
+
+## Useful Ontoly Capabilities
+
+- `ExplainArchitecture` for repository and package topology
+- `FindDependencies` for dependency trees and direct consumers
+- `ImpactAnalysis` for refactor blast radius
+- `TraceExecution` for request, route, and call-flow tracing
+- `FindConfigurationUsage` for configuration and environment variable usage
+- `FrameworkReport` for detected framework concepts such as modules, controllers, providers, and routes
+- `FindDeadCode` for unreachable or unused graph regions
+
+## Answer Shape
+
+When answering, include:
+
+- the direct answer
+- graph evidence
+- confidence
+- diagnostics or caveats
+- fallback source inspection only if needed
+
+Example:
+
+```text
+AuthController handles authentication.
+
+Evidence:
+- node: class:src/auth/auth.controller.ts:AuthController
+- route edges: HANDLES POST /login, POST /logout
+- dependency edges: USES AuthService, JwtService
+
+Confidence: high, because the graph has controller, route, and dependency edges with source locations.
+```
+
+## Fallback Rules
+
+- If the graph is missing, build it first when allowed.
+- If graph validation fails, report the failure and use source search only to verify the affected area.
+- If multiple nodes match the same name, ask for disambiguation or show the candidates with package/module context.
+- If the requested concept is not in the graph, return `NOT_FOUND` with the closest graph evidence instead of inventing an answer.
+

skills/ontoly-software-graph/metadata.json

@@ -0,0 +1,29 @@
+{
+  "name": "ontoly-software-graph",
+  "version": "1.0.0",
+  "description": "Use Ontoly's deterministic Software Graph and MCP capabilities for architecture review, request tracing, dependency analysis, configuration lookup, and impact analysis.",
+  "author": "0xsarwagya",
+  "dependencies": {
+    "node": ">=20"
+  },
+  "platforms": ["macOS", "Linux", "Windows"],
+  "tags": [
+    "ontoly",
+    "software-graph",
+    "mcp",
+    "typescript",
+    "architecture",
+    "dependency-analysis",
+    "impact-analysis",
+    "codebase-understanding"
+  ],
+  "triggers": [
+    "Ontoly",
+    "Software Graph",
+    "trace request",
+    "impact analysis",
+    "dependency graph",
+    "architecture review",
+    "configuration audit"
+  ]
+}