# AI use cases for Genart Skill (camilleroux/genart-skill) Repo: [camilleroux/genart-skill](https://github.com/camilleroux/genart-skill) | Use cases: 4 | Page: https://dedupedai.com/use-cases/camilleroux-genart-skill/ A Claude Code plugin that teaches Claude generative art practice and ships two scripts that check a sketch for determinism and render seeds, sheets and rarity censuses. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Check a sketch is deterministic before minting: https://dedupedai.com/use-cases/camilleroux-genart-skill/check-sketch-determinism-before-minting/ - Render one hash to a PNG and adjust the sketch: https://dedupedai.com/use-cases/camilleroux-genart-skill/render-one-seed-to-close-visual-loop/ - Measure the rarity your trait table actually produces: https://dedupedai.com/use-cases/camilleroux-genart-skill/measure-rarity-of-trait-table/ - Review a whole edition on one sheet, then export PNGs: https://dedupedai.com/use-cases/camilleroux-genart-skill/review-edition-as-contact-sheet-and-batch-pngs/ ## How to check a sketch is deterministic before minting with Genart Skill Repo: [camilleroux/genart-skill](https://github.com/camilleroux/genart-skill) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Run repeatability, distinctness, global state and feature checks on a hash-seeded sketch. **Who it is for:** A generative artist about to mint or release a hash-seeded piece on an onchain platform who wants proof the same hash gives the same render. **The problem:** A sketch that uses Math.random() or leaks state between runs can look fine and still produce different outputs for the same token hash. **The outcome:** A printed report of ok and FAIL lines for repeatability, distinctness, global state and features, so the artist knows what was and was not verified. ### Steps 1. `/plugin marketplace add camilleroux/genart-skill` 2. `/plugin install genart@camilleroux-genart` 3. `npm i -D playwright && npx playwright install chromium` (inferred, check the README) 4. `node "$CLAUDE_PLUGIN_ROOT/scripts/check.mjs" myproject` (inferred, check the README) ### Prompt to paste into your agent ```text Using the genart plugin from camilleroux/genart-skill, run node "$CLAUDE_PLUGIN_ROOT/scripts/check.mjs" on the sketch folder . Expect a report with repeatability, distinctness, global state and features sections ending in 'all checks passed' or a list of failures. Verify that the repeatability lines show the same render hash across the repeated runs for each seed before saying the sketch is deterministic. ``` ### Caveat The checks prove same-machine reproducibility, perceptual stability and feature stability only, not cross-machine determinism, which the plugin says it cannot prove. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/camilleroux-genart-skill/check-sketch-determinism-before-minting/ ## How to render one hash to a PNG and adjust the sketch with Genart Skill Repo: [camilleroux/genart-skill](https://github.com/camilleroux/genart-skill) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Render a single token hash to a PNG so Claude can look at the result and edit the sketch. **Who it is for:** A generative artist iterating on a sketch with Claude Code who wants to see how one specific seed looks after each change. **The problem:** Judging a generative sketch by reading code is slow, and the artist needs to see the actual composition for a given hash. **The outcome:** One PNG for the chosen hash that Claude can open, inspect and use to decide the next edit. ### Steps 1. `/plugin marketplace add camilleroux/genart-skill` 2. `/plugin install genart@camilleroux-genart` 3. `npm i -D playwright && npx playwright install chromium` (inferred, check the README) 4. `node "$CLAUDE_PLUGIN_ROOT/scripts/render.mjs" myproject --hash 0x…` ### Prompt to paste into your agent ```text Using the genart plugin from camilleroux/genart-skill, run node "$CLAUDE_PLUGIN_ROOT/scripts/render.mjs" --hash <0x-token-hash> to produce one PNG. Expect a single PNG file for that hash. Verify the PNG opens and matches the hash you passed before editing the sketch again. ``` ### Caveat The script needs Playwright installed in your project, and the plugin itself stays dependency-free. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/camilleroux-genart-skill/render-one-seed-to-close-visual-loop/ ## How to measure the rarity your trait table actually produces with Genart Skill Repo: [camilleroux/genart-skill](https://github.com/camilleroux/genart-skill) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Run a census over thousands of hashes to see how often each trait appears. **Who it is for:** A generative artist designing traits and rarity tables for an edition who wants real frequencies instead of guessed weights. **The problem:** Rarity weights written by hand often do not match what the code produces across an edition. **The outcome:** A measured rarity report from a census run, showing what the trait table produces across the sample size. ### Steps 1. `/plugin marketplace add camilleroux/genart-skill` 2. `/plugin install genart@camilleroux-genart` 3. `npm i -D playwright && npx playwright install chromium` (inferred, check the README) 4. `node "$CLAUDE_PLUGIN_ROOT/scripts/render.mjs" myproject --census 5000` ### Prompt to paste into your agent ```text Using the genart plugin from camilleroux/genart-skill, run node "$CLAUDE_PLUGIN_ROOT/scripts/render.mjs" --census 5000 and report the measured trait frequencies. Expect a rarity summary for the 5000 sampled hashes. Verify that the measured percentages are compared against the intended weights in the trait table before changing any weights. ``` ### Caveat The census measures what the code produces for the sampled hashes, so it shows the rarity of this sketch, not any guarantee for a live edition. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/camilleroux-genart-skill/measure-rarity-of-trait-table/ ## How to review a whole edition on one sheet, then export PNGs with Genart Skill Repo: [camilleroux/genart-skill](https://github.com/camilleroux/genart-skill) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Make a contact sheet of many seeds, then batch-export individual PNGs for review or delivery. **Who it is for:** A generative artist preparing a full edition who wants to review many seeds at once and then export individual files. **The problem:** Checking an edition one hash at a time misses how the pieces look together, and exporting files by hand is tedious. **The outcome:** A grid image showing many seeds on one sheet, plus individual PNG files for each batch seed. ### Steps 1. `/plugin marketplace add camilleroux/genart-skill` 2. `/plugin install genart@camilleroux-genart` 3. `npm i -D playwright && npx playwright install chromium` (inferred, check the README) 4. `node "$CLAUDE_PLUGIN_ROOT/scripts/render.mjs" myproject --grid 50` 5. `node "$CLAUDE_PLUGIN_ROOT/scripts/render.mjs" myproject --batch 50` ### Prompt to paste into your agent ```text Using the genart plugin from camilleroux/genart-skill, run render.mjs on with --grid 50 and then with --batch 50. Expect one contact sheet image and fifty individual PNG files. Verify that the grid shows fifty distinct seeds before the batch files are used for review. ``` ### Caveat The grid and batch outputs show what the sketch renders, so they cannot confirm print or pen plotter output, which the README does not describe in detail. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/camilleroux-genart-skill/review-edition-as-contact-sheet-and-batch-pngs/