Menu
Scaffold a scheduled task (cron or interval) in app/tasks. Use when asked to run something periodically, on a timer, or on a schedule.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Scaffold a new slash command in app/commands. Use when asked to add or create a slash command, a subcommand, or an autocomplete option.
Scaffold a persistent button, modal, or select-menu handler in app/components. Use when asked to handle a component interaction by customId.
Scaffold a right-click context-menu command (on a user or message) in app/context. Use when asked to add a context-menu or right-click command.
Scaffold a gateway event handler in app/events. Use when asked to react to a Discord event like GuildMemberAdd, MessageCreate, or Ready.
Scaffold a new text (prefix) command in app/prefixes, e.g. ".ping". Use when asked to add a message-based or prefix command.
Scaffold an HTTP API route in app/routes on the Elysia server. Use when asked to add an HTTP endpoint, REST route, or webhook receiver.
| name | new-task |
| description | Scaffold a scheduled task (cron or interval) in app/tasks. Use when asked to run something periodically, on a timer, or on a schedule. |
.agents/context.md (Scheduled tasks section) if not already loaded.app/tasks/<name>.ts.BaseTask, decorated with @Cron("...") or @Interval("..."); implement execute().import { Interval, BaseTask } from "engine";
@Interval("5m")
export default class Presence extends BaseTask {
async execute() {
this.client.user?.setActivity("with decorators");
}
}
@Interval("30s" | "5m" | 60000): fires on a setInterval; accepts a lib/duration string or raw milliseconds.@Cron("*/5 * * * *"): a 5-field cron expression (engine/cron.ts), evaluated once per minute.setInterval or setTimeout loop inside a command, event, or index.ts; declare a task instead.this.client is injected. Wrap unsafe work in try/catch; a crashing task must not take the process down.bun run shard every shard runs its own copy. Guard with a shard check if the work must run once cluster-wide.Run bun run typecheck and bun run lint. Tasks are live on next boot; no deploy step.