aii-schedule-job

Creates scheduled recurring jobs that resolve tools by category from a client registry.

Updated Aug 21, 2026
One-click install
npx skills add https://github.com/maestroventures/aiintegrator --skill aii-schedule-job-maestroventures
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: aii-schedule-job
Source: https://github.com/maestroventures/aiintegrator/tree/main/plugins/aii-blueprint/skills/aii-schedule-job
Command: npx skills add https://github.com/maestroventures/aiintegrator --skill aii-schedule-job-maestroventures

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Scheduled jobs built from memory hardcode tool addresses that silently go stale, so a dead job and a job that never fired look identical and nobody notices until work goes missing. This Skill enforces one sanctioned process for creating any recurring or unattended job so it stays resolvable, auditable, and provably alive. ## Core Features & Use Cases - Category-based tool resolution: Every job opens with a gate that resolves tools by category and action from the client's own registry, never by hardcoded connector name or address. - Runtime and reach discipline: Forces a deliberate hosted-vs-local choice, records what the job must touch (tenant-store-only vs operator-files), and stores the schedule as a tenant-owned row claimable by any awake executor. - Proof before live status: Requires one real run confirming tool resolution, landed work, and a written heartbeat before a job may be called live. - Use Case: When a user says "sync the CRM every morning" or "run this sweep nightly," the Skill names the job by outcome, writes its opening gate, registers it in the client registry, and verifies the first run. ## Quick Start Ask the assistant to schedule a nightly sync job and have it resolve the required tools by category from the client registry before proving the first run.

Frequently Asked Questions about aii-schedule-job

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I schedule a recurring job that runs unattended?

Use this Skill whenever work should repeat on a cadence, such as nightly syncs or weekly sweeps. It names the job by outcome, picks a hosted or local runtime, writes an opening gate that resolves tools by category, and proves the job with one real run before calling it live.

Why should scheduled jobs resolve tools by category instead of name?

Tool addresses are facts about one installation on one day and silently change when connectors re-provision. Resolving by category and action against the client's registry returns a candidate set of live addresses, so the job keeps working when any single address goes stale.

Should a scheduled job run hosted in the cloud or locally?

Default to hosted, because a local job stops when the operator's machine is off and can hang silently on permission prompts. Choose local only when the job must touch operator files, and record that as reach equals operator-files with a hosted route owed.

What happens when the account running a scheduled job hits its usage limit?

The schedule lives as a row in the tenant's store, not inside one account, so any awake executor on any signed-in platform can claim it atomically. A lease ensures a dead executor releases the job back to the queue instead of stranding it.

When should I not use a scheduled job?

Do not schedule one-off work in the current session, steps a human must run on their own machine, or polling inside an open conversation. Do the one-off work immediately and offer the schedule afterward, or route human-run steps through a run-command skill.

Why does a scheduled job need a heartbeat and a proof run?

Without an alive stamp written at claim time, a job that dies in its opening gate is indistinguishable from one that never fired. One real run confirming tool resolution, landed work, and written state is required before the job may be reported as live.