local_rag

To retrieve local documentation chunks, provide a repo_name and a natural-language query to the local_rag Skill. It searches local knowledge bases for math and scientific Python packages like sympy and scipy, returning relevant snippets. An optional top_k parameter controls the exact number of chunks retrieved for grounded question answering.

Finding specific usage examples in local documentation requires a query-focused chunk search. By specifying the package as repo_name and your question as the query, you retrieve the exact documentation passages needed, enabling downstream reasoning to cite and follow package-specific evidence without relying solely on web search.

Yes, local package documentation retrieval applies directly to grounded question answering and coding support. It fetches relevant documentation snippets from locally available knowledge bases for math and science-focused Python packages, ensuring your downstream reasoning has package-specific doc evidence to cite and follow.

You control the number of retrieved documentation chunks using the optional top_k parameter. When you query local knowledge bases for math and scientific Python packages, top_k determines exactly how many relevant snippets are returned for your natural-language question, bounding the retrieved evidence.

The main limitation is that local documentation retrieval is restricted to locally available knowledge bases for math- and science-focused Python packages. It also operates with a bounded timeout and requires specific repo_name and query inputs, meaning it cannot fetch external web evidence or unsupported package documentation.