菜单
Create and apply a Drizzle schema change. Use when adding or altering tables, columns, or indexes in @packages/db, or when asked for a migration.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Add a new shared workspace package under packages/*. Use when creating a new @packages/<name>, whether a bundled library other workspaces import or a build-only script package that runs during a build.
Orient in this repo: which file to edit for a change, how a change ripples across the stack, and how to search the code. Use at the start of a task in an unfamiliar area, or before a cross-cutting change.
Prefer Bun-native APIs, else Node built-ins with the node: prefix. Use when importing a Node built-in (fs, path, child_process, crypto, os, util), reading or writing files, spawning a process, or choosing between a Bun and a Node API.
Icebox a raised-but-undecided concern instead of forcing a plan-or-dismiss call: record it with no verdict so the context survives. Use when a review, PR, audit, or eval surfaces something real-maybe that should not be scheduled or closed yet, or when the user says to icebox or park an item.
Add a typed Hono API endpoint or WebSocket route: router, OpenAPI docs, validation envelope, and RPC client wiring. Use when adding or modifying routes in api/hono.
Start, restart, and verify the ZeroStarter dev stack. `bun run dev` serves portless named `.localhost` URLs (branch-prefixed in a worktree); resolve them with `bunx portless get`. Use when asked to run the app, when the API returns NOT_FOUND for routes that exist in source, or before browser testing.
| name | db-migration |
| description | Create and apply a Drizzle schema change. Use when adding or altering tables, columns, or indexes in @packages/db, or when asked for a migration. |
PostgreSQL + Drizzle ORM. Schema in packages/db/src/schema/, migrations in packages/db/drizzle/. Schema, SQL, and snapshot travel together in one PR.
packages/db/src/schema/<name>.ts, then export it from index.ts: export * from "@/schema/<name>". Every new table must be exported there or it never reaches a migration.auth.ts for tables, relations, and indexes; waitlist.ts for a minimal non-auth table.text primary keys (.$defaultFn(() => crypto.randomUUID()) on non-auth tables), timestamp("created_at").defaultNow().notNull(), snake_case columns, onDelete: "cascade" on FKs, an index() on every FK column.bun run db:generate
Read the generated packages/db/drizzle/NNNN_*.sql. Done when that SQL, its meta/NNNN_snapshot.json, and a new meta/_journal.json entry all appear and the SQL matches the schema edit.
bun run db:migrate
Local / ad-hoc only. The API build auto-applies pending migrations on production and canary deploys (.github/scripts/migrate-on-deploy.ts, gated on VERCEL_ENV/VERCEL_GIT_COMMIT_REF; PR previews skipped), so a migration merged to canary applies itself on the next deploy.
bunx turbo run build --filter=@packages/db
The API consumes @packages/db's built dist. If dev is running and the API imports the new table, restart dev entirely: bun --hot does not pick up new files or exports reliably (see the dev skill).
bun run db:studio
POSTGRES_URL comes from the root .env.