一键导入
这个仓库中的 skills
Use when the task needs a SQL function or virtual table from the gosqlite ext/ catalog — regexp, uuid, hash, ipaddr, zorder, stats, unicode, encode, text, fuzzy, decimal, money, time, eval, csv, lines, statement, closure, pivot, rtree, series, array, bloom, spellfix1, fileio, blobio.
Use when debugging surprising behaviour with gosqlite, or as a pre-ship checklist — the gotchas around per-conn hooks, the connection pool, vec quirks, bind limits, and what is not supported.
Use when the task needs encryption at rest, storing a database compressed at rest, page-checksum corruption detection, an in-memory database with isolation, or serving a database from an embed.FS / byte buffer in a Go app with gosqlite.
Use when using gorm.io/gorm with gosqlite — the gosqlite.org/gorm dialector, a CGo-free drop-in for glebarez/sqlite and gorm.io/driver/sqlite, shipped as its own module.
Use when switching an existing Go project to gosqlite from mattn/go-sqlite3, modernc.org/sqlite, glebarez/sqlite, gorm.io/driver/sqlite, or ncruces/go-sqlite3.
Use when opening a SQLite database in Go with gosqlite.org, choosing the "sqlite" vs "sqlite3" driver name, or deciding which sub-package to reach for. The starting point for any task using this driver.
Use when storing large, growable, or randomly-writable byte objects (files, uploads, streamed content) in SQLite with gosqlite — the blobstore package, plus when to reach for the fixed-size Conn.OpenBlob / ext/blobio instead.
Use when backing a writable SQLite database with arbitrary Go storage (object store, fault injector, custom in-memory) in a Go app with gosqlite — implementing the vfs.VFS / vfs.File interfaces and vfs.Register.
Use when the task wants an ORM or declarative models on top of gosqlite — LiteORM adds first-class native vector / full-text / hybrid (RRF) search and encryption on the gosqlite SQLite driver. Also when porting gorm code to LiteORM, or when one app needs the same data layer across SQLite and Postgres/MySQL/SQL Server.
Use when adding keyword, full-text, or FTS5/BM25 search to a Go app with gosqlite — the typed fts.Index[K, V] API, tokenizers (incl. custom Go tokenizers), query builder, ranking, snippet/highlight.
Use when combining vector (semantic) and full-text (lexical) search results into one ranking in a Go app with gosqlite — Reciprocal Rank Fusion via the fusion package.
Use when adding semantic, similarity, embedding, or KNN vector search to a Go app with gosqlite — the typed vec.Table API over sqlite-vec, distance metrics, encodings, quantization, and metadata filters.