Back Office · office.temerarii.xyz
One asset, all the way in — composition, the wireframe + storyboard, the output format stack, and the template, all read from the SAME content-index record. The expected output matches what /media surfaces for this post.
post longform-W43-Sunkind longformweek W43date 2026-10-25campaign longform-youtubepillar performancebeat asset videoduration 241.9sground redscenes 10

Checklist the per-video bar — engine/sim

98.6/100
plain languagevo coverageno dead airuniquenesscaption fitcompletenesscleanliness
quantitative quality · weights learn from your reviews (engine.sim.memory review longform-W43-Sun good|bad)
⚠ 1 flag(s) — not yet ship-ready: custom_element_dup · see docs/strategy/VIDEO-CHECKLIST.md

Composition comp · template family · expected output

composition LongFormChaptersfamily / template LongFormChapters
9:16 Reelpending1:1 Squarepending16:9 Widepending9:16 4Kpending1:1 4Kpending16:9 4KpendingGIF (SMS)pending
render pending — silent master not yet on disk
expected output: 0/7 rendered — same matrix the /media preview surfaces for this asset.

Composition layer × scene 10 scenes · 241.9s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d
open0–28.4sspatial-parallaxicon·white-knockout♪ bed_in
Everyone wants the finished video, the clean photo, the live stream that runs without a hitch. Nobody wants to talk about the boring part that makes all of it possible: the plan. This week the calendar says the proof you are real comes from a plan you can see. So today we open the content strategy itself, the map that decides what gets made and why, and we run the whole thing in public with The Big T-M.
on-screen: The plan comes before the proof
expected on screen: red ground · dodeca hero in the shared Signal Field · Mensor leads · node-graph · spatial-parallax · icon·white-knockout logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape dodecaground redtreatment white-knockoutmotion spatial-parallaxpower summoninstrument summon→The plan comes before the proof
2s2
first render · fix pending
NumberedList
templatecustom_element_dup
teach28.4–53.2skinetic-buildicon·white-knockout♪ node_lock
The first move is to stop scattering ideas across notes apps and group chats. Put your whole content plan in one plain file. We use a single JSON file that lists every week, every day, every asset. The model reads that one file and knows the entire year. You do not manage a hundred posts. You manage one source of truth, and everything downstream is built from it.
on-screen: Start with one source file
expected on screen: red ground · a NumberedList panel over a dimmed Signal Field · Mensor leads · node-graph · kinetic-build · icon·white-knockout logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape dodecaground redtreatment white-knockoutmotion kinetic-buildpower morphinstrument morph+laser→Start with one source filecurate items, nodes
3s3
matches intent
ChecklistCard
template
teach53.2–78.4skinetic-buildicon·liquid-chrome♪ node_lock
Next, wire that file into the model directly. Point Claude Code at the file with a read tool. Now the agent is not guessing what week it is or what the topic should be. It opens the file, finds this week, and pulls the campaign, the headline, and the services we are teaching. The plan tells the work what to do, instead of you re-explaining it every single time.
on-screen: Let the model read your calendar
expected on screen: red ground · a ChecklistCard panel over a dimmed Signal Field · Mensor leads · node-graph · kinetic-build · icon·liquid-chrome logo · caption bottom-left
spec (the prompt): comp_id ChecklistCardvisual node-graphshape dodecaground redtreatment liquid-chromemotion kinetic-buildpower morphinstrument morph+laser→Let the model read your calendarcurate items, nodes
4s4
matches intent
StackTrace
template
teach78.4–102.2skinetic-buildicon·white-knockout♪ node_lock
A content strategy is not a vibe. It is a list. This week the pillar is multimedia, and under it sit eight concrete services: strategy, live streaming, photography, screenwriting, drone, interviews, advertising, and event coverage. We assign each day a different one so no two videos repeat. That rule lives in the plan, not in your head, so the agent can enforce it for you.
on-screen: One pillar, eight real services
expected on screen: red ground · a StackTrace panel over a dimmed Signal Field · Mensor leads · node-graph · kinetic-build · icon·white-knockout logo · caption bottom-left
spec (the prompt): comp_id StackTracevisual node-graphshape dodecaground redtreatment white-knockoutmotion kinetic-buildpower morphinstrument morph+laser→One pillar, eight real servicescurate errMsg, errType, fix, frames, nodes
5s5
matches intent
BlueprintGrid
template
teach102.2–127.0skinetic-buildicon·liquid-chrome♪ node_lock
Here is the step you can copy today. Take your week and give each day a single distinct angle. Write it down as data, one row per day. When the model builds the scripts, it reads that row and stays on topic. You stop getting seven videos that all say the same thing, because the plan made each day pick a different lane before a word was written.
on-screen: Map each day to a job
expected on screen: red ground · a BlueprintGrid panel over a dimmed Signal Field · Mensor leads · node-graph · kinetic-build · icon·liquid-chrome logo · caption bottom-left
spec (the prompt): comp_id BlueprintGridvisual node-graphshape dodecaground redtreatment liquid-chromemotion kinetic-buildpower morphinstrument morph+laser→Map each day to a jobcurate nodes, rows
6s6
matches intent
CheatSheet
template
teach127.0–149.3skinetic-buildicon·white-knockout♪ node_lock
Every long-form here follows a beat list: open, teach, teach, proof, resolve. That list is also in the file. The model authors exactly that many chapters, in that order. So the structure is not improvised at midnight. It is a contract. You can change the shape of a video by editing four words in a file, and the next render obeys.
on-screen: Beats decide the shape
expected on screen: red ground · a CheatSheet panel over a dimmed Signal Field · Mensor leads · node-graph · kinetic-build · icon·white-knockout logo · caption bottom-left
spec (the prompt): comp_id CheatSheetvisual node-graphshape dodecaground redtreatment white-knockoutmotion kinetic-buildpower morphinstrument morph+laser→Beats decide the shapecurate nodes, points, uses
7s7
matches intent
StepFlow
template
teach149.3–170.20000000000002skinetic-buildicon·liquid-chrome♪ node_lock
Once the plan is data, you can show it. We turn the same file into a public office page, every week, every asset, laid out so anyone can inspect it. No screenshots, no slides. The page is generated from the file. When the plan changes, the page changes, because they are the same thing wearing two outfits.
on-screen: Render the plan into pages
expected on screen: red ground · a StepFlow panel over a dimmed Signal Field · Mensor leads · node-graph · kinetic-build · icon·liquid-chrome logo · caption bottom-left
spec (the prompt): comp_id StepFlowvisual node-graphshape dodecaground redtreatment liquid-chromemotion kinetic-buildpower morphinstrument morph+laser→Render the plan into pagescurate nodes, steps
8s8
matches intent
NodeGraphCard
template
teach170.2–194.0skinetic-buildicon·white-knockout♪ node_lock
The last teaching move is the one most people skip. Run a checker over your own plan. We have a small simulation pass that reads every script and flags repeated lines, missing topics, or empty hype words. The model grades its own homework before a human ever watches. That is how you ship a calendar of work without reading all of it line by line.
on-screen: Let the machine check itself
expected on screen: red ground · a NodeGraphCard panel over a dimmed Signal Field · Mensor leads · node-graph · kinetic-build · icon·white-knockout logo · caption bottom-left
spec (the prompt): comp_id NodeGraphCardvisual node-graphshape dodecaground redtreatment white-knockoutmotion kinetic-buildpower morphinstrument morph+laser→Let the machine check itselfcurate hub, nodes
9s9
matches intent
ComparisonTable
template
proof194.0–218.5sreceipts-counticon·liquid-chrome♪ node_lock
Here is the honest proof. This video you are watching was built from that one file. The topic, the structure, the order of these chapters, all of it came out of the plan we just described. We did not write this by hand and then pretend a machine did it. The plan is the maker. You can see the whole thing at office dot temerarii dot xyz.
on-screen: You are watching it run now
expected on screen: red ground · a ComparisonTable panel over a dimmed Signal Field · Mensor leads · receipts · receipts-count · icon·liquid-chrome logo · caption bottom-left
spec (the prompt): comp_id ComparisonTablevisual receiptsshape dodecaground redtreatment liquid-chromemotion receipts-countpower receiptsinstrument spotlight→You are watching it run nowcurate colA, colB, rows, statsLabels
10s10
matches intent
shared field
signature-3d
resolve218.5–241.9scoalescenceicon·white-knockout♪ bed_out
So that is the move. Put your content plan in one file, wire a model to read it, give each day a distinct job, and let a checker keep you honest. The proof everyone wants is just the last step of a plan they never saw. Go look at ours, take the parts that fit, and point your own machine at your own file.
on-screen: See the plan, then copy it
expected on screen: red ground · dodeca hero in the shared Signal Field · Mensor leads · coalescence · coalescence · icon·white-knockout logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape dodecaground redtreatment white-knockoutmotion coalescencepower coalescenceinstrument coalescence→See the plan, then copy it

Format stack 1 aspects · same scenes[], re-cropped

16:9
1920×1080
X/Twitter · YouTube · LinkedIn video

Channels 2 destinations

YouTubeBlog

Social captions supplemental published copy · per channel (comp_id level)

youtubeHow to plan a content week so the proof is built, not promised This week is the plan behind the proof. Before you can show you are real, you need a plan that decides what gets made, in what order, and how it gets checked. This video shows how to plan a content week so the proof is built, not promised. The method: Start the plan as one file. The campaign, the days, the service or method each day teaches, and the beats for each piece. The plan is the source every script and asset reads from, so nothing gets made off the cuff. Give each day a different job. One concrete method per day, named out loud, so the week does not blur into the same safe message repeated seven times. The plan forces variety by assigning different facts to different days. Set the rule the model writes against: deep teach one move, no hype, plain language, no fake numbers. The model drafts the schedule and the scripts, you read them like an editor and send back anything that sounds fake. Wire the checks into the plan. A banned word list, a color check, a logo only rule, all run before anything ships. The plan is not just what to make, it is what has to pass. The whole point is that the plan absorbs the pressure. When the schedule, the beats, and the checks are decided in advance, making the work is execution, not invention. The boring planning is what lets the proof get built every single week without a scramble. Every week of our calendar runs on this exact plan, in public. See it at office.temerarii.xyz. Keywords: content planning, content calendar, editorial plan, AI content, quality checks, automation.

Cross-links every lens is a view on this one record