Publish 3d Flag

---

name: gotchi-3d-custom-render description: Render custom Aavegotchi 3D images from arbitrary trait and wearable combinations. Use when the user describes a synthetic or hypothetical gotchi look in plain language, asks for an outfit preview, headshot, or full-body image, and is not asking for an existing onchain token by tokenId or inventory URL.

---

gotchi-3d-custom-render

Use this skill when the user wants a custom 3D gotchi render from selected traits, wearable IDs, wearable names, or a named preset.

Plain-language requests should also route here, for example:

"render an ETH gotchi with common eyes, aagent hat, shades, shirt, headset, and pistol"

"show me a custom gotchi portrait on a cream background"

"make a gotchi outfit preview with wizard hat and blue wand"

"give me a headshot of a hypothetical gotchi with shades and a fedora"

Do not wait for the user to explicitly say gotchi-3d-custom-render.

What this skill does

writes a structured render request JSON

uses the hosted Aavegotchi renderer by default for official-style output

falls back to the local Unity batch worker when hosted assets are unavailable

returns:

- full-body PNG - headshot PNG - manifest JSON

supports friendlier presets, slot aliases, themed backgrounds, and render-mode selection

wearable flags accept either numeric IDs or quoted wearable names

supports plain-language outfit requests after routing picks this skill

Constraints

requires a local Unity editor installation

hosted mode matches the existing Aavegotchi renderer look much more closely, but may ignore custom local-only pose behavior

Unity fallback is still useful for hypothetical combinations or local debugging

hosted mode can run on non-Unity hosts as long as node, jq, and npm install have been run in the repo root so the bundled PNG background compositor is available when a non-transparent background is requested

intended for local development first, then AAi/Telegram integration through hosted mode

no environment variables are required for normal hosted rendering

Routing Notes

Prefer this skill when the user describes a look, outfit, traits, wearables, background, portrait, headshot, or hypothetical gotchi.

Prefer aavegotchi-3d-renderer instead when the user gives a tokenId or inventory URL for an existing gotchi.

If the user gives both a tokenId and custom outfit instructions, clarify whether they want the real token render or a hypothetical custom loadout.

Entry points

main wrapper: scripts/render-custom-gotchi.sh

preset gallery helper: scripts/render-preset-gallery.sh

direct Unity runner: scripts/run-unity-render.sh

Unity project: unity/GotchiCustomRenderer

Quick usage

From the skill root:

bash scripts/render-custom-gotchi.sh \
  --preset aagent-eth

Default mode is hosted. Use --render-mode auto if you specifically want hosted-first with Unity fallback, or --render-mode unity for local-only rendering.

Outputs land in:

Renders/-full.png

Renders/-headshot.png

Renders/-manifest.json

Input contract

Supported flags:

--preset

--slug

--haunt-id

--collateral

--eye-shape

--eye-color

--skin-id

--background

--pose

--body

--face

--eyes

--head

--pet

--hand-left

--hand-right

--left-hand

--right-hand

--slot

--wearables

--print-presets

--write-only

--render-mode

--hosted-only

--unity-only

--find-wearable

Wearable flags can take either:

numeric IDs like --head 59

quoted names like --head 'Aagent Fedora Hat'

Friendly Examples

bash scripts/render-custom-gotchi.sh --find-wearable aagent
bash scripts/render-custom-gotchi.sh --preset portrait-eth
bash scripts/render-custom-gotchi.sh --preset aagent-eth
bash scripts/render-custom-gotchi.sh --preset aagent-eth --render-mode auto
bash scripts/render-custom-gotchi.sh --body 'Aagent Shirt' --face 'Aagent Headset' --eyes 'Aagent Shades' --head 'Aagent Fedora Hat' --hand-right 'Aagent Pistol'
bash scripts/render-custom-gotchi.sh --wearables 'head=56,hand-right=58' --bg gotchi-radio --pose hero
bash scripts/render-custom-gotchi.sh --slot hat=56 --slot weapon=58 --background transparent
bash scripts/render-preset-gallery.sh
bash scripts/render-preset-gallery.sh --preset aagent-eth --preset portrait-eth --gallery-name quicklook
bash scripts/render-preset-gallery.sh --preset aagent-eth --preset portrait-eth --gallery-name quicklook --skip-failures
bash scripts/render-preset-gallery.sh --preset aagent-eth --gallery-name quicklook --open

Natural-language request examples this skill should handle:

"render a custom ETH gotchi with common eyes and an aagent pistol"

"show me a gotchi with aagent hat, shades, shirt, and headset"

"make a portrait gotchi on a studio cream background"

"preview this outfit on an ETH gotchi with common eyes"

If you need to inspect the request shape or adjust it manually, read references/request-schema.md. If you need the preset list or alias vocabulary, read references/presets.md.

When to read extra files

read references/request-schema.md when you need the exact JSON payload layout

read references/presets.md when you need preset names, background aliases, or friendlier slot names

read references/wearables.tsv when you need the built-in wearable name catalog used on non-Unity hosts

read unity/GotchiCustomRenderer/Assets/Editor/GotchiCustomRenderCLI.cs only when changing the Unity render behavior

requires a local Unity editor installation

hosted mode matches the existing Aavegotchi renderer look much more closely, but may ignore custom local-only pose behavior

Unity fallback is still useful for hypothetical combinations or local debugging

hosted mode can run on non-Unity hosts as long as node, jq, and npm install have been run in the repo root so the bundled PNG background compositor is available when a non-transparent background is requested

intended for local development first, then AAi/Telegram integration through hosted mode

no environment variables are required for normal hosted rendering