docs-retrieval

Local-first doc retrieval checks your on-device ai-docs store before web search, loading targeted pages and consolidating context for faster lookup. Start by checking ai-docs/libraries/_index.toon to identify relevant docs, then load specific library indexes and consolidate pages into a single context block.

Targeted loading retrieves concise page summaries instead of full documents, reducing token usage and speeding up context consolidation. Load specific library indexes, select relevant pages, and merge them into a focused context block for your sub-agent.

Docs-retrieval follows the docs-management protocol for structured workflows. Load ai-docs library indexes, identify pages matching your task, consolidate context, and pass it to sub-agents for feature implementation, debugging, or library exploration with reproducible results.

Local-first retrieval falls back to web search when local docs are insufficient. Check your ai-docs store first using the library index, then query the web if you don't find what you need, ensuring you exhaust on-device sources before external lookup.

Yes, docs-retrieval applies across multiple docs by loading and consolidating context from different library indexes. Use the ai-docs/_index.toon to discover available libraries, then load targeted pages from each relevant set and merge them into unified context.