Andromeda

Skill NovateStudioGit/novate-studio-skills/paid-growth/andromeda

57 agent skills for Claude Code — creative production, paid growth, copywriting, ecommerce, email marketing & knowledge ops. By Novate Studio.

Install

npx -y skills add NovateStudioGit/novate-studio-skills --skill andromeda

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

0 stars0 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.

What its author says it does

Copied from the file, not written here

ANDROMEDA - the creative engine. Turns a brand's products + personas + angles into a diverse batch of distinct creative CONCEPTS (persona x angle x offer), sized to budget, then briefs them out across the production pipeline. Enforces the rules that make creative actually scale: the production split (25 campaign-shoot / 40 UGC / 20 organic-repurpose / 15 static), UGC rules (no-talk for global scale, size-on-screen to cut returns, trending audio, multi-product for longevity), desire-vs-painpoint mix (60-70/30-40 toward pain points), evergreen-first creative, and the brand-vs-direct-response balance. Every batch = distinct concepts across distinct cohorts, never variations of one idea. Triggered by `/andromeda` or natural language like "build me a creative batch", "what concepts should we test", "plan this month's ads", "ideate ad concepts for [brand]", "creative engine", "concept matrix". Feeds the production skills (`/model`, `/style`, `/ugc`, `/product-ugc`, `/packshot`, `/static-ads`); concepts get placed by TRINITY (Meta) and ARGUS (Google). NOT a renderer itself - it produces the concepts + briefs.

SKILL.md

3.2 KB, as published. Nobody here has run it

ANDROMEDA - the creative engine

The brain that decides what creative to make before anything gets produced. It generates a wide, diverse batch of distinct concepts (never variations of one idea), sizes the batch to budget, and briefs them across the production pipeline. Promoted from a principle to a skill on 2026-06-03; mirrors the "meta creative engine" in the ExampleBrand Notion.

The brain (read before reasoning)

reference/concept-matrix.md - persona x angle x offer; generating at scale; desire-vs-painpoint; jewellery/ExampleBrand adaptation.
reference/formats-and-production.md - production split, UGC rules, evergreen-first, brand-vs-DR, the AI-creative-returns caution.

Pipeline

0. SCOPE     brand? budget / TRINITY spend tier (sets batch width)? a product set or the full catalogue?
1. INPUTS    pull personas (brand doc or survey), products (catalogue), and ideate angles
2. MATRIX    build persona x angle x offer concepts; dedupe to genuinely distinct ideas/cohorts
3. WIDTH     size the batch to budget + spend tier; rank concepts by fit/expected value
4. FORMATS   assign formats per concept against the production split; flag evergreen vs seasonal
5. BRIEF     output concept briefs (persona + angle + offer + formats + hook direction)
6. HANDOFF   route to /model /style /ugc /product-ugc /packshot /static-ads; tag concepts for KPI

House rules

Distinct concepts across distinct cohorts, never one idea re-skinned. Width sized to budget.
Tag every concept so produced ads can be KPI'd by concept (the creative feedback loop).
Evergreen-first for jewellery (winners can run for years - build the compounding library).
AI for context (mood/lifestyle/models), real photography or faithful packshots for the actual product in a visual vertical (returns risk). No em/en dashes (memory no-em-en-dashes).

Handoffs

Production: /model, /style, /ugc, /product-ugc, /packshot, /static-ads. Placement: TRINITY (Meta) and ARGUS (Google Shopping). Economics: MIDAS. Competitors: /ad-spy.