skill-publish-scheduler · Batch Publish Scheduler

A batch scheduling skill for timed publishing: it manages a "content × platform × publish time" schedule — importing plans, viewing the queue, computing due items, and dispatching each due item to the right platform publishing skill, then marking statuses back. The scheduling logic is pure standard-library scripting; it never drives a browser or calls a publishing API itself.
"Watchdog" on-time triggering depends on an upper-layer cron job calling it periodically; the script only computes due items per run.
Example invocation: "Schedule these three posts for Douyin and Xiaohongshu at 8pm Wednesday and Friday."
Full brief
Positioning
skill-publish-scheduler is the scheduling and status layer: it manages a "content × platform × publish time" schedule, dispatches due items to each platform's publishing skill, and marks statuses back. It handles scheduling and state only — real publishing is delegated to the platform publishers.
Core capabilities
- Schedule import: accepts CSV (headers must include
publish_at, platform, content; optionaltitle, note) or equivalent JSON. - Queue view & due computation:
listviews the queue;due/runcompute what's due now and the dispatch plan. --nowdry rehearsal: overrides the current time to preview what would publish at a given moment, without affecting real time.- Per-platform delegation: each due item is handed to its platform's publisher skill (skill-xhs-publisher / skill-douyin-upload / skill-wechat-publisher / skill-bilibili-upload / cross-platform-publish), with content adapted per platform before publishing.
- Status marking:
markrecords each item's outcome; can pair with skill-publish-notify for result pushes and skill-publish-log for records. - Pure-script scheduling: scheduling logic in pure stdlib, no resident process.
Workflow
importthe schedulelistto view the queuedue/runto get due items and the dispatch plan- Delegate each due item to its platform publisher for real publishing
markeach item's status back
Inputs & outputs
| Input | Notes |
|---|---|
| Schedule | Required — CSV or equivalent JSON, headers must include publish_at, platform, content |
Output: queue view, due items, dispatch plan, status-mark confirmations.
Boundaries with adjacent skills
- Platform publisher skills (skill-xhs-publisher / skill-douyin-upload / skill-wechat-publisher / skill-bilibili-upload): drive browsers/APIs for real publishing (environment-dependent). This skill only manages scheduling and status — it never publishes directly.
Fit
- Distributing multiple posts to platforms on a fixed timetable
- Simultaneous multi-platform publishing at one moment
- Rehearsing a publish plan for a future moment
- Scenarios needing unified tracking of publish statuses
Before you start
- Prepare a schedule with
publish_at, platform, content; useYYYY-MM-DD HH:MMfor times; multiple platforms at the same moment go due together. - Timed triggering depends on an upper layer (OpenClaw cron or manual periodic runs) — the script is stateless due-computation plus a persisted queue, not a daemon.
- Always
markafter each publish to avoid duplicates.
skill-publish-scheduler is part of the Aiglade Skill library. Invoke it from the Aiglade chat box in plain language.