gooseworks-ai/goose-skills / skills/video/render-flat-vector-explainer
render-flat-vector-explainer
VerifiedAssemble the FREE steps of the flat-vector-explainer video format — a flat-illustration host character (who, where, tone, voice and music come from the recipe choices) walks a countable N-step product routine, one step per beat, and Remotion composites every chip/numeral/tagline/slate/CTA as an animated DOM overlay ON TOP of the Kling i2v character clips (text is NEVER baked into a keyframe — i2v warps type), the closing 'N products' grid is a PIL composite of the REAL product photos (not AI), full-sentence VO drives word-by-word burned captions over a VO-forward music bed, and the ~50s animated silent master is re-cut to a 30s deliverable FROM the animated master (never a static intermediate). Documentation-grade — ships config.example.json + PIPELINE.md + a README of the free assembly; the paid gen steps (keyframes, Kling i2v, VO, music) are separate capabilities the recipe orchestrates. Use for the flat-vector-explainer format.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 6.2 KB |
| scripts/PIPELINE.md | markdown | 4.4 KB |
| scripts/README.md | markdown | 5.6 KB |
| scripts/config.example.json | config | 10.7 KB |
| skill.meta.json | config | 183 B |
| tests/smoke-test.md | markdown | 1.2 KB |
scripts/PIPELINE.md
View on GitHub# PIPELINE — flat-vector-explainer engine map
This molecule is **documentation-grade**. Rather than re-implement a 14-scene Remotion
app here, it maps every `config.json` field to the **per-project script** that
builds that step of the demo. Those scripts are not shipped here.
Port or rebuild them in the new brand's project folder (Kling i2v + Remotion overlay),
run the pipeline there, then keep the config here as the recipe of record.
**The demo's creative values are examples, not defaults.** The host, setting, tone, narrator
voice and music come from the recipe's `choices` (`host`, `setting`, `tone`,
`narrator_voice`, `music`). The demo (a mid-30s woman at a sunny windowsill, faintly
comedic, "Eryn" VO, lo-fi pop) is one worked example.
## Config field → source script
| `config.json` field | Source script | Phase | Paid? |
|---|---|---|---|
| `character.anchor_prompt`, `character.keyframe_variants`, `scenes[].keyframe_prompt` | `working/gen_keyframes.py` | 1 — character lock (host from `choices.host`) + per-scene keyframes in the chosen setting (nano-banana, flat-vector; re-render a FRESH anchor — never chain a photoreal ref) | **PAID** |
| (clean plates — strip baked text) | `working/clean_plate.py` | 2 — nano-banana edit removes chips/numerals/taglines/badges from character keyframes → clean plates for i2v | **PAID** |
| `scenes[].motion`, `kling.*` | `working/kling_i2v.py` | 3 — Kling v3 standard i2v (`start_image_url`, `generate_audio: false`; 2.5 Turbo is withdrawn) on character scenes only; cfg 0.5 + style-preserving negative + subtle motion; TEST one scene first | **PAID** |
| `scenes[].overlay` (chips, numerals, taglines, slate, CTA) | `working/remotion/` (Remotion project) | 4 — imports each Kling clip as the moving base and composites ALL text as animated DOM ON TOP → animated silent master. Text is NEVER baked into a keyframe | free |
| `product_grid.*` | `working/scripts/build_scene08.py` | 4 — PIL composite of the N real product webps on the brand ground (preserve each aspect; AI duplicates SKUs so this is PIL, not AI) | free |
| `voice.*`, `scenes[].vo` | `working/scripts/render_vo.py` | 5 — ElevenLabs eleven_v3, `text-to-speech/with-timestamps`, voice from `choices.narrator_voice`; full-sentence per-scene VO in the chosen tone + char-level timestamps | **PAID** |
| `music.*` | `working/gen_music.py` | 5 — ElevenLabs music bed in the style from `choices.music` (demo: lo-fi pop), VO-forward loudnorm; skipped if the user chose no music | **PAID** |
| (VO+music mix) | `working/scripts/mix_audio.sh` | 5 — place each VO line at its scene start, duck music under VO (sidechaincompress), `loudnorm I=-15`, VO-forward | free |
| `captions.*` | `working/scripts/build_captions.py` | 5 — word-by-word burned captions from the VO char-timestamps (libass); suppress on slate/grid/CTA scenes | free |
| (silent master assembly) | `working/scripts/build_silent.sh` | 4 — stitch the Remotion-composited scenes into `working/silent-master.mp4` (the ANIMATED master) | free |
| (50s master) | `working/build_master.py` | 5 — mux silent master + mixed audio, burn captions LAST → `finals/master-final.mp4` | free |
| `cutdowns.*` | `working/build_30s.py` | 6 — slice each beat's region OUT of `silent-master.mp4` (the ANIMATED master, NEVER a static intermediate); trim short / slow long (≤1.6×); re-burn scaled captions → `finals/master-final-30s-v1.mp4` | free |
| (QC) | `working/motion_probe.py` | 7 — frame-diff proof of localized motion (+ `/watch` the final) | free |
## The two non-negotiable separations
1. **Motion layer ≠ text layer.** `clean_plate.py` strips text → `kling_i2v.py` animates
the clean plate → `remotion/` composites text as DOM on top. Never `kling_i2v.py` a
keyframe with baked text (i2v warps it, and you can't retime/restyle it).
2. **Real assets ≠ AI assets.** `build_scene08.py` PIL-composites the real product webps.
AI gen duplicates SKUs in a grid — it is only for the character vignettes + backgrounds.
## The cut-down trap
`build_30s.py` MUST slice from `working/silent-master.mp4` (the Kling-animated master),
NOT from `working/segs/` (an earlier static Ken-Burns round). The mtime is the tell.
A finished-looking audio mix can hide frozen characters — always run `motion_probe.py`
frame-diff on a character scene and confirm **localized** face/hand glow (whole-outline
glow = pan-only = wrong source).
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |