project-query-docs-index

Offline documentation search is performed by querying three local SQLite FTS5 indexes covering the workspace clone, global skills store, and local docs, which returns BM25-ranked results without internet access.

BM25 ranking scores document relevance using term frequency and inverse document frequency, and this skill applies it to offline SQLite FTS5 indexes to surface the most relevant skills, rules, and docs for quick discovery.

Yes, offline SQLite FTS5 search supports query operators and scope filters across the workspace, global skills, and local docs indexes to refine BM25-ranked results for precise document discovery.

No, offline BM25 search requires no external services or web access, querying three local SQLite FTS5 indexes entirely offline-first to locate skills, rules, and docs without internet connectivity.

The best way to locate SKILL.md files offline is by querying the workspace clone and global skills store SQLite FTS5 indexes, which returns BM25-ranked top matches for immediate reading and activation of relevant skills.

Limitations of offline SQLite FTS5 search include dependency on local index freshness and confinement to the workspace, global skills, and local docs indexes, meaning new content requires manual index updates before discovery.