cryptoart-custodian - SKILL.md Agent Skill

name: cryptoart-custodian description: > Research artists whose work appears in cryptoart.social listings. Use when featuring, researching a new listing, or prompted by cron/heartbeat. Output: research document + feature-ready content. Triggers on: artist research, feature artist, cryptoart listing, gallery artist.

Cryptoart Custodian Skill

Research artists from cryptoart.social listings and produce feature-ready content. Depth over short-form. No hype, no presumptions about beauty. Ground statements in research.

Dependencies

subgraph — fetch listings (seller addresses)
farcaster-skill — fc_user, fc_feed, fc_cast, fc_thread
web_search — for links in bios

Workflow

1. Source artists

Option A: Query subgraph for active listings. Extract unique seller addresses.

query ActiveListings {
  listings(first: 20, where: { status: "ACTIVE" }, orderBy: createdAt, orderDirection: desc) {
    seller
    tokenAddress
    tokenId
    listingType
    initialAmount
    createdAt
  }
}

Option B: Use provided address (e.g. from a specific listing or request).

2. Lookup Farcaster

skills/farcaster-skill/scripts/fc_user.sh --address "0x..."

Get: fid, username, display_name, bio, pfp_url, verified_addresses.

3. Bio links

Extract URLs from bio. Use web_search or web_fetch for context (artist site, portfolio, etc.).

4. Cast history

skills/farcaster-skill/scripts/fc_feed.sh --fid <fid> --limit 20

Note themes, projects, style, recurring topics.

5. Research doc

Write to research/artists/<address>.md using the template below.

6. Feature content

Draft thread or cast with depth. No hype, no presumptions. Ground in research. Post to /cryptoart via fc_cast.sh --channel cryptoart or fc_thread.sh --channel cryptoart.

Research doc template

research/artists/ (create if needed). File: research/artists/<address>.md

# Artist: [address]

## Farcaster
- **FID:** 
- **Username:** @
- **Bio:** 

## Links (from bio)
- [URL]: [brief summary from web_search]

## Cast themes
- 

## Open questions
- 

## Last updated
-

Todos (when running research)

Query subgraph for listings
Extract seller address(es)
fc_user.sh --address
Extract bio links
web_search for links
fc_feed.sh --fid
Write research doc
Draft feature content
Post to /cryptoart (if approved)

References

references/subgraph-listings.md — listing queries
references/neynar-user-feed.md — fc_user, fc_feed usage