# Agent Scheduled Tasks

> Use when creating, changing, debugging, or proving an agent-owned scheduled task, automation, cron, reminder, or heartbeat.

- Skill: `adsagents/agent-scheduled-tasks` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add adsagents/agent-scheduled-tasks`
- Raw SKILL.md: https://api.skillmd.com/api/skills/adsagents/agent-scheduled-tasks/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: adsagents (https://skillmd.com/u/adsagents)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/adsagents/agent-scheduled-tasks

---


# Agent Scheduled Tasks

1. Classify the job as reminder/heartbeat, auditable execution, or
   consequential execution.
2. Resolve timezone, cadence, destination, stable scope, allowed tools,
   evidence, retry budget, and stop conditions before creating it.
3. Use the actual scheduler interface once, then read back its stored contract.
4. Run now only when supported and record the scheduler plus downstream
   evidence separately.
5. Never put credentials in the prompt, guess permissions, or treat schedule
   creation as execution proof.
6. Treat a matching delivery prepare as the live preflight. Consume inline
   verification first and use the advertised native operation fallback only
   when recovery remains pending; a missing optional selector before mutation
   does not disable read-only schedule evaluation.
7. For explicit Meta candidates, batch up to 50 exact IDs with
   `overview_get_live_configs`; correlate only by `entity_type` and `entity_id`.

Read [scheduled-task-contract.md](scheduled-task-contract.md) only when
designing, creating, updating, debugging, pausing, deleting, or proving a
scheduled task.