How vedats/ragskill-dotnet works
RagSkill-Dotnet is a Claude Code skill that scaffolds a runnable .NET document-Q&A application or integrates its components into an existing app. The generated app’s user workflow is document upload and indexing, followed by question answering with retrieved chunks and clickable citations; its UI also manages documents and persistent chat history. The template and scripts are visible in the tree, while runtime wiring is described by the README’s architecture and feature summary; unsampled implementation details are not expanded into unsupported call edges.
Main entry points: SemanticSearch.cs Semantic search · DataIngestor.cs Document ingestion · Chat.razor Chat page · SKILL.md Skill instructions
Components 21
Skill workflow
- Skill instructions
skills/dotnet-rag-chat/SKILL.md - Scaffold script
skills/dotnet-rag-chat/scripts/scaffold.py - Integration script
skills/dotnet-rag-chat/scripts/integrate.py - Runnable template
skills/dotnet-rag-chat/assets/template
Generated interface
- Chat page
skills/dotnet-rag-chat/assets/template/Components/Pages/Chat/Chat.razor - Documents page
skills/dotnet-rag-chat/assets/template/Components/Pages/Documents.razor - Citation viewer
skills/dotnet-rag-chat/assets/template/Components/Pages/Chat/ChatCitation.razor
RAG runtime
- App registration
skills/dotnet-rag-chat/assets/template/Program.cs - Document ingestion
skills/dotnet-rag-chat/assets/template/Services/Ingestion/DataIngestor.cs - Document reader
skills/dotnet-rag-chat/assets/template/Services/Ingestion/DocumentReader.cs - Index synchronization
skills/dotnet-rag-chat/assets/template/Services/Ingestion/DocumentSyncService.cs - Semantic search
skills/dotnet-rag-chat/assets/template/Services/SemanticSearch.cs - Embedding retries
skills/dotnet-rag-chat/assets/template/Services/RetryingEmbeddingGenerator.cs
Models and storage
- Indexed chunks
skills/dotnet-rag-chat/assets/template/Services/IngestedChunk.cs - Chat history store
skills/dotnet-rag-chat/assets/template/Services/ChatHistoryStore.cs - SQLite vector store
Other components
- Developer
- Claude Code
- Ollama models
- App user
- PDF / Markdown
skills/dotnet-rag-chat/assets/template/wwwroot/Data
Connections 24
- Developer → Claude Code: requests app
- Claude Code → Skill instructions: follows
- Skill instructions → Scaffold script: creates project
- Scaffold script → Runnable template: copies and configures
- Skill instructions → Integration script: integrates project
- Integration script → Runnable template: copies components
- App user → Documents page: uploads and deletes
- Documents page → Document ingestion: indexes documents
- PDF / Markdown → Document reader: read and chunk
- Document ingestion → Document reader: uses reader
- App registration → Index synchronization: starts indexing
- Index synchronization → Document ingestion: synchronizes
- Document ingestion → Embedding retries: embeds chunks
- Embedding retries → Ollama models: requests embeddings
- Document ingestion → SQLite vector store: stores vectors
- App user → Chat page: asks question
- Chat page → Semantic search: search tool
- Semantic search → Embedding retries: embeds query
- Semantic search → SQLite vector store: retrieves top chunks
- Semantic search → Ollama models: requests chat response
- Chat page → Chat history store: loads and saves
- Semantic search → Indexed chunks: uses retrieved content
- Chat page → Citation viewer: renders citations
- Citation viewer → App user: opens source quote