Skip to content
Back to skills

Visual Story 6 Production

ASecurity

STEP 6 of 8 in the sequential visual-story pipeline. Scene production. Builds the STYLE LOCK block, establishes master shots, generates every mini-scene's coverage attaching the locked character/location/prop references, uses the "Change to..." technique for sequential shots, and propagates continuity/weather across a scene. Run AFTER visual-story-5-exposition (reference sheets built). Next step: visual-story-7-platform-qc.

  • 2 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added October 9, 2026
ai-agents

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned October 9, 2026

npx -y skills add teen-boop/AiProductionPipeline --skill visual-story-6-production --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Visual Story 6 Production?

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

Security grade badge for Visual Story 6 Production
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/teen-boop-visual-story-6-production/badge)](https://www.skillsdirectory.com/skills/teen-boop-visual-story-6-production)

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: visual-story-6-production
description: STEP 6 of 8 in the sequential visual-story pipeline. Scene production. Builds the STYLE LOCK block, establishes master shots, generates every mini-scene's coverage attaching the locked character/location/prop references, uses the "Change to..." technique for sequential shots, and propagates continuity/weather across a scene. Run AFTER visual-story-5-exposition (reference sheets built). Next step: visual-story-7-platform-qc.
---

# Visual story — Step 4: Scene production

**Step 4 of 5.** Before this: `visual-story-4-references` (approved reference sheets for every
character/location/prop must exist). After this: `visual-story-7-platform-qc`.

Full mechanics, prompt skeletons, and worked examples are in `references/prompt-engineering.md` —
read it; this SKILL.md is the checklist.

## Build the STYLE LOCK block first (once per project)

After the style references and first approved masters exist, write a STYLE LOCK block (photographic
mode, light, depth of field, atmosphere, palette, film/technical, aspect ratio) and paste it
**verbatim** into every production prompt. Never re-derive the look from memory per-prompt — that's
how drift creeps in over a long shot list. Template in `references/prompt-engineering.md`.

## Scene continuity lock (run before writing the scene's prompts)
Run `scene-continuity-lock` for every multi-shot scene: build the scene state sheet (who sits where on
which furniture, props and their states, light, camera axis and mirroring for reverse shots), check each
prompt against the previous and the next one, paste the CONTINUITY block verbatim into every prompt,
and QC each generated frame against it. Background wall per shot comes from `location-room-map`.

## Master shot + continuity per scene

- The first approved wide/establishing shot of a location is that scene's **master** — attach it
  (plus the location reference sheet from step 3) to every other shot in the scene, not just
  re-described in text.
- **Weather/physical-state propagation:** if the master shows rain/snow/mud, every character
  entering that space afterward shows visible evidence of it (wet hair, damp fabric) — stated
  explicitly, not left to the model.
- Every locked visual detail (a stain, a torn sleeve, a prop, a hairstyle state) is copied
  **verbatim** across every prompt in the same mini-scene.

## Attach references on every generation

For each shot, attach: the scene's master shot → each in-frame character's sheet + real face photo +
approved portrait → each in-frame location/prop reference sheet. One face photo per person in a
multi-character frame. (Character-lock attachment order detail lives in step 3's
`character-lock.md`.)

## "Change to..." for sequential coverage within a mini-scene

To build the next beat of a mini-scene (wide→close-up, gaze shift, next action, or relocating an
established look), use an already-approved shot as a reference and write a prompt starting with
"Change to [the next beat]: keep/same [restated costume + identity details]." Always also attach the
character's real face photo, even alongside the base image. This depicts the next beat of ACTION in
the same continuous scene — not an unrelated alternate. Full technique + the three variant types
(reframe / redirect gaze / relocate) in `references/prompt-engineering.md`.

## Prompt series per mini-scene

Each mini-scene gets one anchor prompt plus additional prompts scaled to how many discrete actions
happen in it — a simple establishing beat needs 1–2; a multi-action sequence needs 5–6+, one per
beat.

## When done

Every scene's coverage is generated against the locked style + references. **Next:
`visual-story-7-platform-qc` — platform operation notes and the QC pass before delivery.**

Files in this skill

  • SKILL.md3.7 KB
  • references/prompt-engineering.md8 KB

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.