slopfiles
The skill pack we use to produce The Slop Files series. Brief becomes 6 first-frame/last-frame video prompts in minutes.
Not published as a public repository yet
Cinematic series production, scripted as a skill. Built for our flagship investigative show, open-sourced so anyone can produce serialized video at scale.
The Problem
Producing one good video is hard. Producing five in a series, with consistent aesthetic, voice, and visual continuity, is brutal. Most series die after episode 2 because production overhead compounds.
What It Does
- 01Brief becomes 6 scene specifications with first/last frame prompts
- 02Visual continuity locked across episodes via shared style config
- 03Per-episode metadata and social cut prompts auto-generated
Built for The Slop Files
slopfiles is the production pipeline we use to make The Slop Files, WhyStrohm's flagship investigative series about AI slop, brand voice drift, and the founder content bottleneck.
Each episode follows the same structure: 6 scenes, each with a first frame and a last frame for fal.ai's first-frame/last-frame video generation. The skill enforces visual continuity (lighting, color palette, atmospheric details) across all episodes so the series feels coherent.
What you get per episode
slopfiles-ep01/
├── BRIEF.md
├── stills/ ← 11 source frames
├── scripts/animate-scenes.py ← fal.ai pipeline
├── narration/ ← per-scene VO
└── output/
├── clips/ ← 6 animated clips
└── final/ ← composited episode
Production cost per episode: roughly $2-5 in fal.ai spend, 3-4 hours of human time.
Why open-source the production engine
We built slopfiles for ourselves. We open-sourced it because the bottleneck for serialized video isn't a tool, it's a system. Anyone can use Kling. Almost nobody can ship 5 episodes with locked visual continuity, narration alignment, and per-episode metadata in a week.
If you want to run your own investigative series, different topic, different aesthetic, fork the skill, swap the style config, run it.
Install
git clone https://github.com/whystrohm/slopfiles.git ~/.claude/skills/slopfiles
Full docs and the production brief template on GitHub →
How It Composes
Sits next to shotkit in the production layer. shotkit handles general video; slopfiles handles serialized investigative content. Both consume brand-config.json from media-tsunami.
Related Skills
shotkit
Pre-production for founder-led video at scale. Brief becomes storyboard, shot specs, and per-generator prompts in minutes.
Install →media-tsunami
The empirical layer. Extracts brand voice as executable code, cadence, vocabulary, forbidden words, exemplar sentences, serialized as a CLAUDE.md any LLM can load.
Install →