# The weekly batch loop and the gate checklist (free sample)

A free sample from Content System Kit by PaperChaseWebb. Works in Grok Bot, Claude, ChatGPT, Claude Code or Codex.

Agents can do most of the production work on your content. The part that goes wrong is publishing: work that nobody read going live at 3 AM. This sample gives you the one rule, the weekly loop built around it and the checklist a person runs before anything goes out.

## The one rule

**Production may run unattended, publication may not.** Agents can draft, cut, caption and queue overnight. Nothing posts, schedules or spends until a person approves that exact item.

## The weekly batch loop

1. **Study (Monday, agent).** Read the account's own last week: what performed, what flopped, one pattern to repeat. One screen, not a data dump.
2. **Plan (Monday, agent drafts, you pick).** Three to five concepts, each tied to something you made, decided or learned yourself. You pick or kill each one.
3. **Produce (Tuesday to Thursday, agents).** Draft each picked piece in your format, with your real logo and assets. Captions drafted, not posted.
4. **Approve (Friday, you).** The whole week arrives as one batch. Approve, edit or hold each item. Only approved items are scheduled.

Keep the account connected through the platform's own tools or your scheduler, under your login. Do not hand account credentials to an unknown connector so an agent can manage it.

## The gate checklist

- Every claim in the piece has a source or is removed
- Every face, voice and track is yours, licensed, or has written consent
- Captions match the audio word for word
- Nothing private is visible (screens, addresses, other people's messages)
- The piece still sounds like you after the agent pass

## Check prompt

"Run this gate checklist on the draft below. List every claim without a source, every face, voice or track without a licence or consent, caption mismatches and anything private on screen. Do not fix anything yourself; list it for me."

## In the full kit

Audience filters, one owned format and a material bank, a 90-post test with day-90 rules set in advance, an editor queue with a board and edit briefs, transcript-first cleanup with an honest scope, a packaging watch on public data only, teardown without replication, a craft seat map, a cadence plan, an engagement honesty audit, a failure table, a worked example, a rules block and 14 prompts.

Sources for this sample: Jayant (@jayantcreates.ai), https://www.instagram.com/reel/DebrTUOj50f/ and Mino Lee (@minolee.mp4), https://www.instagram.com/reel/DejCX63y2LM/. Rebuilt in our own words; their results are their claims.

Made by PaperChaseWebb at paperchasewebb.ai.