docs-seeker

To fetch up-to-date API references, use a prioritized retrieval hierarchy that checks llms.txt on official docs, Context7 mirrors, and GitMCP. This retrieves authoritative, current documentation with version-awareness to prevent confidently wrong code generation from stale model memory.

Version-aware documentation retrieval prevents wrong-major answers by checking package manifests and lockfiles to fetch version-specific docs. This ensures the agent retrieves exact function signatures and required fields mapped to the correct library version, avoiding outdated API behavior.

The best way to get third-party framework docs into context safely is using WebFetch for a few small pages and fanning out to parallel subagents for large pages. This strategy returns condensed answers with source URLs and fetch metadata without exceeding context limits.

Yes, you can use llms.txt and Context7 to find exact function signatures for Stripe or FastAPI. The retrieval hierarchy prioritizes these sources to surface exact signatures, required fields, and relevant notes directly from authoritative documentation into the agent's context.

You need to fetch official documentation when projects rely on external libraries whose APIs change over time. Fetching current documentation ensures you retrieve authoritative, version-specific behavior and feature-scoped details instead of guessing from stale model knowledge.

Yes, docs-seeker works with WebFetch for reading documentation pages. It uses WebFetch for a few small pages and fans out to parallel subagents for many or large pages, returning condensed answers with source URLs and fetch metadata to manage context safely.