knowledge-sources

To sync external documentation into a searchable knowledge base, this Skill connects to GitHub repositories, local files, URLs, and APIs, pulling content and normalizing it into a unified Markdown format for efficient retrieval.

You can search synced external content without a dedicated vector database by using file-based search mechanisms. The Skill leverages standard tools like grep to efficiently query across the normalized Markdown knowledge cache.

Incremental syncing for external knowledge sources works by detecting and pulling only the changed files during synchronization. This approach saves time and bandwidth when updating content from GitHub, local directories, or APIs.

The best way to make diverse file types searchable together is through content normalization. This Skill converts content from GitHub, local file systems, and web URLs into a unified, searchable Markdown format within a single knowledge cache.

Content synchronization from GitHub and web URLs supports automated daily updates. You can configure a repository like 'my-org/my-docs' to sync daily into the knowledge cache, ensuring your agents access the latest external information.

A limitation of using grep for searching synced external content is that it relies on file-based text matching rather than semantic search. This means queries must use exact keywords present within the normalized Markdown files to retrieve relevant information.