
Design the posting rhythm for an automated social operation — daily slots anchored to one timezone, auto-fill versus manually-scheduled slot types, per-channel timing adaptation, and a queue-depth target the agent loop maintains. Use when someone asks "when should we post", "set up our posting schedule", "how many posts per day", or "design the queue for our scheduler". Do NOT use for deciding content themes or pillars — that is strategy/calendar territory; for connecting and probing the scheduler, use social-stack-bootstrap; for keeping the queue actually filled, use scheduling-agent-loop.
---
name: Posting Cadence Designer
description: Design the posting rhythm for an automated social operation — daily slots anchored to one timezone, auto-fill versus manually-scheduled slot types, per-channel timing adaptation, and a queue-depth target the agent loop maintains. Use when someone asks "when should we post", "set up our posting schedule", "how many posts per day", or "design the queue for our scheduler". Do NOT use for deciding content themes or pillars — that is strategy/calendar territory; for connecting and probing the scheduler, use social-stack-bootstrap; for keeping the queue actually filled, use scheduling-agent-loop.
---
# Posting Cadence Designer
Cadence is infrastructure, not preference. The output is a CADENCE.md that a scheduling agent can execute against without judgment calls.
## Inputs
STACK.md from social-stack-bootstrap (channels, scheduler capabilities), the audience's primary timezone(s), and the realistic content production capacity per week. Cadence that outruns production capacity fails silently — design to the capacity you have, not the capacity you want.
## Workflow
**1. Anchor to one operating timezone.** All slots are stated in it. Per-channel local-time adaptations are derived from the anchor, never stated independently — two sources of truth for time is how posts land at 3 a.m.
**2. Define the slots.** Start with two or three daily slots (e.g. morning, noon, end-of-day). More slots than distinct daily content ideas produces filler; fewer than two wastes scheduler automation. For each slot record: time, channels it serves, and content weight (primary announcement vs. lighter engagement post).
**3. Classify each slot: auto-fill or manual.** Schedulers treat these differently and their tooling often only surfaces auto-fill slots. Document the blind spot explicitly: "The queue-schedule endpoint only shows auto-fill slots; the <time> slot is manually scheduled and will NOT appear there — absence from the tool is not absence from the schedule." This single note prevents an agent from double-booking or falsely reporting an empty queue.
**4. Set the queue-depth target.** How many days ahead must the queue be filled — typically 7–10 days. Below target is the agent loop's trigger to generate and schedule. Also set a hard ceiling (e.g. 14 days): content scheduled too far out goes stale, and stale scheduled posts are the most common silent failure in automated social (see scheduling-agent-loop's stale-content audit).
**5. Adapt per channel, minimally.** Only add per-channel exceptions that earn their complexity: e.g. a community platform (Reddit) keyed to when its communities are active, not to brand slots. Every exception is one more thing the agent loop must check — default to the shared slots.