feat(api): add pipeline execution, save, and list endpoints

Add pipelines table to SQLite schema with Drizzle migration.
Implement POST /api/v1/pipeline/execute (sequential multi-tool processing),
POST /api/v1/pipeline/save, GET /api/v1/pipeline/list,
DELETE /api/v1/pipeline/:id. Pipeline execution validates all tool IDs
and settings before processing, chains output of each step as input
to the next.

apps/api/src/db/schema.ts

@@ -44,3 +44,11 @@ export const apiKeys = sqliteTable("api_keys", {
   createdAt: integer("created_at", { mode: "timestamp" }).notNull().$defaultFn(() => new Date()),
   lastUsedAt: integer("last_used_at", { mode: "timestamp" }),
 });
+
+export const pipelines = sqliteTable("pipelines", {
+  id: text("id").primaryKey(),
+  name: text("name").notNull(),
+  description: text("description"),
+  steps: text("steps").notNull(), // JSON array of { toolId, settings }
+  createdAt: integer("created_at", { mode: "timestamp" }).notNull().$defaultFn(() => new Date()),
+});