Why I built it
I built the Weekly Content Planner to carry selected, source-linked project activity from the Weekly Builder Update into a separate content-planning draft for the following week.
Internal reports contain useful progress stories, but the builder still needs to choose what is accurate, appropriate to share and supported by real photos or video. This workflow brings those opportunities together for review.
From weekly activity to a review draft
- Weekly Builder Update
- Content Opportunities
- Safety Review
- Content Planner
- Builder Review
The handoff begins after the weekly report drafts have been prepared and logged. It uses the same run’s project and completed-activity records, with a separate planner for each project. It does not copy the private weekly report into the planner.
I use fixed, reviewed templates for a limited set of recognised activities. The current build has no AI-provider dependency. Unknown activity wording is held or excluded rather than interpreted freely.
What the builder can review
Suggested topics, formats, hooks and homeowner value, linked to their source activity.
Practical photo or video requests when usable assets are missing.
Safety status, approval status and notes beside each candidate.
Suggested content dates for the following week, not construction commitments.
There is no fixed posting quota. Quiet weeks do not need filler. Missing assets remain requests, not invented photographs or claims that a past activity can be recreated safely.
Safety checks and builder decisions
Safety review excludes private or sensitive material and explicitly unverified activity. Missing verification or publication permission keeps a candidate under review; a recorded completion status alone is not independent proof of the work.
Each planner is scoped to one builder and project. A project-week reservation prevents a matching replay from creating another draft. Changed content or uncertain delivery is held for reconciliation.
The builder decides what becomes public. Candidates can be approved, changed, skipped or queried. Source checks, actual assets and rights, builder delivery and approval recording stay human-controlled. There is no automatic social publishing. Builder delivery and publishing remain manual.
What I built and tested
The October 5, 2026 build evidence records 40 local tests and 35 n8n logic tests passing, covering source boundaries, privacy, builder and project isolation, missing assets, duplicate replay and uncertain-result handling.
One Gmail coordinator review draft was created and read back successfully. Duplicate replay skipped another draft. The Weekly Builder Update handoff was checked separately using synthetic source records.
These checks demonstrate the review-mode build. The screenshot shows its implementation structure; the saved tests establish the results described here.
