Build generative media
Skill hiteshbandhu/skills-i-use/skills/ai-engineer-talks/build-generative-media
Builds and ships generative image/video products—FLUX, ComfyUI workflows, Veo, fal platforms, Luma-scale launches, and API-first creator UX. Use when working on diffusion pipelines, Comfy graphs, video models, generative media infra, or viral consumer media launches.From its SKILL.md
npx -y skills add hiteshbandhu/skills-i-use --skill build-generative-mediaAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
- runs commandsInstructs the agent to run 2 commands, including `cp -r skills/build-generative-media ~/.cursor/skills/` and 1 more.
SKILL.md
2.0 KB, 432 tokens by cl100k_base, as published. Nobody here has run it
Build generative media
Action playbook from nine AI Engineer generative media talks. Do not summarize — pick a workflow theme.
Supporting files: workflows.md · source-index.md
Optional: ./skill-outputs/build-generative-media/
Step 0 — Pick workflow
What is the user trying to do?
├─ Choose image model (FLUX, open research) → image-models
├─ ComfyUI reproducible node pipelines → comfy-workflows
├─ Viral consumer launch / GPU scale (Luma) → scale-launch
├─ Creator API UX (Replicate-style) → api-ux
├─ Multi-model platform routing (fal) → media-platform
├─ Google Veo / GenMedia enterprise chains → google-video
├─ Diffusion training/serving at scale (DeepMind) → training-scale
└─ Exploration mindset for new modalities → curiosity-lab
Install
cp -r skills/build-generative-media ~/.cursor/skills/
cp -r skills/build-generative-media ~/.codex/skills/
Source: playlists/generative-media-ai-engineer/.
Cross-cutting rules
| Rule | Source |
|---|---|
| Embed workflow JSON in outputs for reproducibility | [src-002 @ 0:01:58] |
| Rehearse 10x scale before viral marketing | [src-003 @ 0:01:37] |
| Video compute ≠ image—budget tokens separately | [src-007 @ 0:04:52] |
| Enterprise gen media needs brand/safety defaults | [src-006 @ 0:15:04] |
| Demoable, legible APIs beat opaque pipelines | [src-004 @ 0:01:15] |
Output
Name theme; artifacts under ./skill-outputs/build-generative-media/ when requested.
What ships with it: 3 files
4.9 KB alongside SKILL.md
- README.md446 B
- source-index.md1.4 KB
- workflows.md3.1 KB
Gives 0 of the 12 instructions most media documents skills give in 432 tokens
Counted across 99 of the 124 authors here whose files we hold, read 2026-09-06
- Define three to five content pillarsin 7 of 99, across 5 files
- Tag every finding as verified, medium, or assumedin 5 of 99, across 3 files
- Respond to comments within one hourin 5 of 99
- Adapt content format to each platformin 5 of 99
- Choose platforms where the audience already spends timein 4 of 99, across 2 files
- Cap promotional content at ten percentin 4 of 99, across 2 files
- Answer product questions within two hoursin 4 of 99, across 2 files
- Acknowledge complaints publicly, resolve them privatelyin 4 of 99, across 2 files
- Batch-create posts to keep cadence consistentin 4 of 99, across 2 files
- Fan out across free media sources in one queryin 3 of 99, across 1 file
- Use Pexels or Pixabay for modern short clipsin 3 of 99, across 1 file
- Extract historical shots from archival filmsin 3 of 99, across 1 file
Said here and by no other author read
- pick a workflow theme first
- route the user goal to one workflow
- embed workflow JSON in outputs for reproducibility
- rehearse tenfold scale before viral launch
- budget video compute separately from image
- set brand and safety defaults for enterprise media
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.