Skip to content
Back to skills

Schedule Workflow

ASecurity

How to author and inspect durable schedules from agent tools

  • 21 stars
  • 0 votes
  • 0 copies
  • 3 views
  • Added September 2, 2026
ai-agentsreact

Security analysis

A100/100

Scanned September 2, 2026

npx -y skills add lukacf/meerkat --skill schedule-workflow --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Schedule Workflow?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Schedule Workflow
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/lukacf-schedule-workflow/badge)](https://www.skillsdirectory.com/skills/lukacf-schedule-workflow)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: schedule-workflow
description: How to author and inspect durable schedules from agent tools
requires_capabilities: [schedule]
---

# Schedule Workflow

Use Schedule when something should happen later, repeatedly, or on a wall-clock
calendar. Schedules own time and delivery. They do not own live work state.

## Operating Rules

- Create schedules with `meerkat_schedule_create` when the user asks for a
  reminder, recurrence, follow-up, monitor, wakeup, or routine automation.
- Choose the smallest trigger that matches the request: `once` for one future
  instant, `interval` for fixed cadence, `calendar` for wall-clock recurrence
  in a named timezone.
- For the current agent session, use the agent-facing session target
  `{"target_kind":"session","type":"current_session","action":{"type":"prompt","prompt":"Check in"}}`.
  The host resolves it to a persisted durable target.
- Use `resumable_session` only when you have the exact durable `session_id`,
  and `materialize_on_demand_session` when the schedule must create a session
  from an explicit build spec on first fire.
- The defaults are `misfire_policy: {"type":"skip"}`,
  `overlap_policy: "skip_if_running"`, and
  `missing_target_policy: "mark_misfired"`. Change them only when the user asks
  for bounded catch-up, concurrency, or different missing-target behavior.
- Inspect with `meerkat_schedule_get`, `meerkat_schedule_list`, and
  `meerkat_schedule_occurrences` before creating duplicates.
- Pause for temporary suspension, resume to reactivate, update to change future
  behavior, and delete only when the schedule should stop permanently. Delete
  is irreversible, but historical occurrences remain queryable.

## Boundaries

- Use WorkGraph for pending work, dependencies, claims, and evidence.
- Use Schedule for time. A scheduled prompt may ask an agent to inspect
  WorkGraph, but Schedule should not duplicate WorkGraph readiness logic.
- Treat an occurrence claim and delivery outcome as Schedule authority only.
  They do not authorize a WorkGraph claim or prove that the underlying shared
  work completed.
- Do not infer successful delivery from a due time. Inspect occurrence phase;
  pending, claimed, dispatching, awaiting completion, completed, skipped,
  misfired, superseded, and delivery failed are distinct states.
- Use memory for recalled knowledge, not for future wakeups.
- Use builtin tasks for private scratch items that do not need scheduled
  delivery or shared durable coordination.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.