konradre/video-production-skills / skills/film-preprod
film-preprod
VerifiedPre-production of a SHORT FILM or a MUSIC VIDEO: the veto map from whoever judges it, the story written end-first (the seven steps in seconds, a subworld per beat), premises ranked twice, the hero shot generated first, the track map measured before the gen map, the shared pre-production core (partitions on continuity, the script diff) and contest hygiene. Use when a film or a music video is briefed, a premise or format must be chosen, a beat sheet must be written or checked, a track must be mapped, or contest rules must become a checklist. Triggers — "story beats", "the seven steps", "the premise", "which format, live action or animation", "the hero shot", "self-revelation", "track map", "cut to the drop", "the jury", "the veto map", "the contest rules". Not for an ad campaign's brief or cost plan — use ad-spot-preprod. Not for the reference set — use video-refs-continuity. Not for the prompt itself — use video-prompt-dialects. Not for the cut — use video-edit-edl.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 10.9 KB |
| references/CONTEST-HYGIENE.md | markdown | 2.2 KB |
| references/HERO-SHOT-AND-FORMAT.md | markdown | 3.4 KB |
| references/JURY-AND-PREMISE.md | markdown | 3.7 KB |
| references/MUSIC-VIDEO.md | markdown | 15.0 KB |
| references/STORY-METHOD.md | markdown | 5.2 KB |
| scripts/arc_check.pyexecutable | script | 4.3 KB |
| scripts/beat_calc.pyexecutable | script | 4.0 KB |
| scripts/premise_rank.pyexecutable | script | 3.1 KB |
| scripts/track_map.pyexecutable | script | 4.9 KB |
references/JURY-AND-PREMISE.md
View on GitHub# Whoever judges the film decides the veto map — then the premise is chosen twice
## 1. Derive the veto map from the judges (or the audience, or the client)
Research each judge's own work and stated taste; find the archetype their careers converge on, and
the one beat that resolves it (for one panel: the ego that discovers its world was authored and must
choose meaning anyway — the anagnorisis, resolved to a single shot: a face under a push-in whose
speed the performance dictates). Then write **two tables before anything is generated**:
- **the veto map** — vetoes are asymmetric: one strong aversion outweighs three mild positives; treat
the column as constraints, not preferences (a tech demo with a thin story · over-lighting, hard
backlight, showy camera, nothing to cut to · exposition that explains the mystery, irony, smugness ·
a sanitised resolution · any screening failure);
- **the compliance table** against the rule set (runtime floor, aspect, subtitles, watermark, no real
likeness, no political/religious statement, all-AI audio, on-platform generation) — each row ✅ / ⚠ /
⬜ with where it is enforced.
The convergence claim is a synthesis — no judge says "here is what I would reward" — so it is
recorded as the method's own inference, and the vetoes (which ARE sourced) carry the weight.
## 2. Blind competitor premises
Generate N alternatives with a **different model family**, each run receiving the identical brief
(the judge dossiers, the archetype, the veto map, the prohibitions, the production limits) and **one
assigned divergence axis** (domestic with no fantastical element · landscape and season · work and
machinery · an invented tradition · the one speculative run) — a single prompt asking for five ideas
returns five variations of one idea. Ban the default metaphors as literal elements; cap the laziest
route (amnesia / simulation / false reality) at one run — it reads as derivative to exactly the judges
who made those films. Do not feed the runs your own premise.
Read the set for what it settles: when three of five resolve to the same shape, they corroborate the
archetype AND they are not alternatives to each other — compare only the materially different bets.
## 3. Rank twice — story and execution — and let them disagree
**Premise selection is a pipeline decision, not only a creative one.** Rank the candidates on how
likely each is to survive the pipeline, separately from which is the best film:
- a world already understood to be made of paint absorbs texture crawl, reconstituting edges and
over-smooth surfaces as diegetic behaviour; a domestic-realism premise has the tightest error bars
that exist (everyone is an expert on kitchens and old hands) and may need legible handwriting;
- night with one motivated source is the most forgiving lighting; landscape is where the models are
strongest; repetitive mechanical motion, water, animals, scale relationships and layered voices
resolving into a younger self are live risks;
- the hero shot needs a real fallback (two clips cut on a caught breath) or the premise is exposed.
`premise_rank.py` prints both rankings; when they disagree, **the disagreement IS the decision** and
goes to the operator as such ("most likely to be executed well is not the same as most likely to
win"). Record each premise's objection (a premise about an AI authoring a world, submitted to an AI
company's festival, trips a judge's "don't editorialise about AI" veto) beside its rank.
## 4. Lock, then publish
When a contest rewards early publication AND binds continuity, lock the premise first (a day of
work), then publish; never publish a placeholder to reserve a date. Everything published becomes
public with its prompts — assume every prompt will be read.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |