laniameda-gallery-query

Use this skill to read from the laniameda.gallery vault. Browse by pillar, semantic search by natural language, retrieve a specific asset with its prompt, or download images to a local tmp folder.

Counterpart to laniameda-gallery-ingest (which writes).

When to trigger

• "find me that car prompt", "get me the cinematic portrait", "show me what's in the gallery"
• "search the gallery for scroll animations"
• "download that image so I can work with it"
• "what prompts do I have for Midjourney?"
• "grab the reference image and use it for..."
• Any time the agent needs to pull content FROM the gallery

Runtime env

• CONVEX_URL or NEXT_PUBLIC_CONVEX_URL — required
• KB_OWNER_USER_ID — required for scope=mine (default: 278674008)