Retail Plugin
Retail AI plugin bundling two installable skills for Google Cloud:
semantic product search and generative image/video virtual try-on.
Follows the Google Agent Plugins v1.0.0 spec (plugin.json +
nested skills/).
Owner: FDE/Blackbelt · POC: tanvisinghal-0105 · License: Apache-2.0
Skills
| Skill | What it builds | Google Cloud services |
|---|---|---|
retail-product-search | Semantic product catalog search / RAG agent for e-commerce, shopping assistants, and catalog discovery. | Vector Search (Gemini Enterprise Agent Platform), BigQuery, embeddings |
retail-virtual-tryon | Virtual try-on agent (image + catwalk video) for clothing, eyewear, jewelry, cosmetics, footwear. Ships a pre-flight product-cutout classifier and configurable safety levels. | Gemini image models, Veo |
Each skill folder contains a SKILL.md (the conversational installer
prompt read by the host assistant) and a README.md (human-oriented
install and surfacing instructions).
Install
Install directly into an AI coding assistant (Claude Code, Antigravity, Codex, Gemini CLI, …):
npx skills add google/adk-recipes --skill retail-product-search
npx skills add google/adk-recipes --skill retail-virtual-tryon
Or, from a cloned copy of this repo, use the per-skill make surface
target to install into ~/.gemini/skills and/or ~/.agents/skills.
See each skill's README.md for the full surfacing matrix.
Deployable companions
The skills bootstrap a working agent locally. If you want the
production-shaped standalone agent — ready for Agent Engine, Cloud Run,
Gemini Enterprise App, and Agent Garden — use the companion contrib/
recipes:
Layout
plugins/retail/
├── plugin.json # plugin manifest (v1.0.0 spec)
└── skills/
├── product-search/
│ ├── SKILL.md # installer prompt (Q-MODE)
│ ├── README.md # human install docs
│ ├── EVAL.yaml # LLM-as-judge + deterministic checks
│ └── … # scripts, references, assets
└── virtual-tryon/
└── … # same shape