Scenario ChatGPT Pet: Update
Overview
An existing pet is its sprite sheet: a PNG or WebP the user downloaded from ChatGPT, a Codex pet
folder (pet.json plus the sheet it names), or a Scenario asset id of a sheet made earlier. That
file is the pet's identity here; there is no pet record to look up. The update never overwrites it:
the result is a new package next to the run, with the same id and name unless the user changes them.
This skill runs the scripts and follows the sheet contract of scenario-chatgpt-pet-create, which
must be installed; run each script by its path in that skill's scripts/ folder, from the user's
working folder, and read the contract before generating a row or writing direction verdicts:
pet_frames.py,
pet_prepare.py,
pet_build.py,
pet_check.py,
pet_preview.py,
pet_package.py, and
the contract. Generation follows that
skill's loop and model order (model_openai-gpt-image-2-5-sunburst, then
model_google-gemini-nano-banana-2-1, then recommend only when both are blocked); connection and
scope: the scenario skill. If a sibling skill named here is missing from your available skills,
load it with skill_get (Scenario MCP) or ask the user to install it
(https://github.com/scenario-labs/skills#install); unattended, proceed from tool schemas and flag
the gap.
Quick reference
Read the sheet first, then pick the smallest change that does the job. Every row below runs in a run
made by pet_prepare.py init --from-split (step 3 of the example): it records the original sheet,
which the build, the check and the package all need. Unattended, the approvals are your own judgment
against the request, recorded in the run, and nothing is installed.
| The user wants |
Do |
Generations |
| To see or check it |
pet_check.py, pet_preview.py contact and gif |
0 |
| A rename or new description |
init, pet_check.py <sheet> --run RUN, then pet_package.py make RUN --sheet <sheet> --name ... |
0 |
| Leftover pixels, color under transparency |
pet_build.py RUN --clean |
0 |
| A row whose frames pop up and down or sideways |
pet_build.py RUN --reregister <row>: re-places the existing frames on one ground line and anchor |
0 |
| One or a few states redone |
init --only <rows> (a look row too), generate those rows, pet_build.py RUN: every other row stays byte for byte |
one per row |
| Look directions on a v1 sheet |
init --version 2 --only look: cardinals, then rows 9 and 10 |
3 |
| A new look (outfit, color, accessory) |
init --change "<the change>": edit the identity, approve it, regenerate every row |
1 + every row |
split refuses anything that is not 1536x1872 or 1536x2288: other art is a new pet, made with
scenario-chatgpt-pet-create using that art as a reference.
Worked example: Pepper's stiff run, then a red scarf for next season
- Get the sheet. The user points at
~/.codex/pets/pepper/: read pet.json (id,
displayName, description, spritesheetPath) and copy the sheet into a work folder. From an
asset id: asset_download with format: "png", then curl -L.
- Read it.
pet_frames.py split pepper.png --out work/split prints the version, a background
key that contrasts with the pet, and pixel_grid (set when the sheet is pixel art with on/off
alpha; the update then keeps that grid and palette).
pet_check.py pepper.png --json-out work/before.json measures it: height per row against idle,
jump lift, look-row drift. pet_preview.py contact and pet_preview.py gif show the current
pet. Ask what should change if the request is vague, naming what the check found.
- Redo the run.
pet_prepare.py init --name Pepper --id pepper --description "<from pet.json>" --from-split work/split --only running-right,running-left --out work/run
(every row of the table starts with an init --from-split like this; --only names the rows to
generate). The match style keeps the sheet's look; each job's references are
references/base.png (the idle frame, enlarged on the key) and references/rows/<row>.png (the
current row). Each job maps onto one model_run: its references uploaded with upload_asset
and passed as referenceImages (an array), its prompt verbatim, its sizes for the model,
background: "opaque" on Sunburst, priced with dry_run first; download the result to the job's
output. When the old row's motion or directions are what is wrong (a jump that never lifts, a
run that does not alternate, a head that turns the wrong way), add --no-current-row to init
so the old strip is not a reference, since the model copies what it is shown, and say what
changes with pet_prepare.py note work/run --jobs <row> --text "..."; a redone look row still
references the kept other half of the sweep. Then
pet_frames.py extract work/run --job running-right, read the report and preview, and mirror it
with --mirror-left or generate running-left the same way.
- Rebuild.
pet_build.py work/run --mirror-left. With the original sheet as the base (recorded
by init), only the rebuilt rows change, scaled to the existing idle row's height, ground line
and anchor; the edge cleanup touches only them.
pet_check.py work/run/final/spritesheet.webp --run work/run must print ok: true; kept look
rows only bring a warning that they were not reviewed, while a redone look row needs fresh
verdicts in qa/direction-semantics.json (format in the contract) or the check fails.
- Show and package.
pet_preview.py gif work/run/final/spritesheet.webp --out-dir work/run/previews, show pet.gif,
upload the sheet and both GIFs into a collection as the create skill does, then
pet_package.py make work/run --id pepper. Ask before pet_package.py install work/run --force,
which backs up the installed Pepper into the run before replacing it.
- The scarf, later. A new look changes every frame, so it is a new generation grounded in the
old pet:
init --name Pepper --id pepper --description "<from pet.json>" --from-split work/split --change "add a red knitted scarf" --out work/scarf
adds an edit job on the identity image. Approve the edited base, then run every row as
scenario-chatgpt-pet-create does from its rows step (look rows included for a v2 sheet, with
fresh direction verdicts). The old sheet still sets the pet's size, so the new pet does not grow
or shrink.
For a v1 to v2 upgrade (init --version 2 --only look), run look-cardinals, look-9 and
look-10 as the create skill describes and write the direction verdicts; the build keeps rows 0-8
and adds the two look rows.
Common mistakes
- Overwriting the user's file or installed pet: work on a copy, and replace an installed Codex pet
only after a yes, with
--force (which backs it up).
- Regenerating the whole pet for one bad row: redo that row; the build copies the others byte for
byte.
- Sending the whole sheet to an image model as the reference: it reads as a grid of small frames and
the model copies the grid. Use
references/base.png and the row strip split made.
- Regenerating to fix what the build can fix: leftover pixels, color under transparency and a row
that drifts are
--clean and --reregister, at no cost.
- Rebuilding without the original sheet: a new row is then sized from nothing;
init --from-split
records the sheet so the build keeps the pet's size.
- Changing the pet's id on a rename: Codex finds the pet by its folder, so keep the id and change
displayName.
- Handing a v2 result to someone uploading on ChatGPT web: give them
package/spritesheet-v1.png.
1---2name: scenario-chatgpt-pet-update3description: "Use when changing an existing ChatGPT pet or Codex pet sprite sheet with Scenario: fixing a broken row, frame, jump or edge, regenerating one animation state, adding the sixteen look directions to a 1536x1872 v1 sheet, giving the pet a new outfit, color or accessory, renaming it in its pet.json, or checking a pet sheet and previewing it as an animated GIF."4license: MIT5metadata:6 source: https://skillmd.com/skills/scenario-labs/scenario-chatgpt-pet-update7---89# Scenario ChatGPT Pet: Update1011## Overview1213An existing pet is its sprite sheet: a PNG or WebP the user downloaded from ChatGPT, a Codex pet14folder (`pet.json` plus the sheet it names), or a Scenario asset id of a sheet made earlier. That15file is the pet's identity here; there is no pet record to look up. The update never overwrites it:16the result is a new package next to the run, with the same id and name unless the user changes them.1718This skill runs the scripts and follows the sheet contract of `scenario-chatgpt-pet-create`, which19must be installed; run each script by its path in that skill's `scripts/` folder, from the user's20working folder, and read the contract before generating a row or writing direction verdicts:21[pet_frames.py](../scenario-chatgpt-pet-create/scripts/pet_frames.py),22[pet_prepare.py](../scenario-chatgpt-pet-create/scripts/pet_prepare.py),23[pet_build.py](../scenario-chatgpt-pet-create/scripts/pet_build.py),24[pet_check.py](../scenario-chatgpt-pet-create/scripts/pet_check.py),25[pet_preview.py](../scenario-chatgpt-pet-create/scripts/pet_preview.py),26[pet_package.py](../scenario-chatgpt-pet-create/scripts/pet_package.py), and27[the contract](../scenario-chatgpt-pet-create/references/sheet-contract.md). Generation follows that28skill's loop and model order (`model_openai-gpt-image-2-5-sunburst`, then29`model_google-gemini-nano-banana-2-1`, then `recommend` only when both are blocked); connection and30scope: the `scenario` skill. If a sibling skill named here is missing from your available skills,31load it with `skill_get` (Scenario MCP) or ask the user to install it32(https://github.com/scenario-labs/skills#install); unattended, proceed from tool schemas and flag33the gap.3435## Quick reference3637Read the sheet first, then pick the smallest change that does the job. Every row below runs in a run38made by `pet_prepare.py init --from-split` (step 3 of the example): it records the original sheet,39which the build, the check and the package all need. Unattended, the approvals are your own judgment40against the request, recorded in the run, and nothing is installed.4142| The user wants | Do | Generations |43| ---------------------------------------------- | ------------------------------------------------------------------------------------------------------------------- | ------------- |44| To see or check it | `pet_check.py`, `pet_preview.py contact` and `gif` | 0 |45| A rename or new description | `init`, `pet_check.py <sheet> --run RUN`, then `pet_package.py make RUN --sheet <sheet> --name ...` | 0 |46| Leftover pixels, color under transparency | `pet_build.py RUN --clean` | 0 |47| A row whose frames pop up and down or sideways | `pet_build.py RUN --reregister <row>`: re-places the existing frames on one ground line and anchor | 0 |48| One or a few states redone | `init --only <rows>` (a look row too), generate those rows, `pet_build.py RUN`: every other row stays byte for byte | one per row |49| Look directions on a v1 sheet | `init --version 2 --only look`: cardinals, then rows 9 and 10 | 3 |50| A new look (outfit, color, accessory) | `init --change "<the change>"`: edit the identity, approve it, regenerate every row | 1 + every row |5152`split` refuses anything that is not 1536x1872 or 1536x2288: other art is a new pet, made with53`scenario-chatgpt-pet-create` using that art as a reference.5455## Worked example: Pepper's stiff run, then a red scarf for next season56571. **Get the sheet.** The user points at `~/.codex/pets/pepper/`: read `pet.json` (`id`,58 `displayName`, `description`, `spritesheetPath`) and copy the sheet into a work folder. From an59 asset id: `asset_download` with `format: "png"`, then `curl -L`.602. **Read it.** `pet_frames.py split pepper.png --out work/split` prints the version, a background61 key that contrasts with the pet, and `pixel_grid` (set when the sheet is pixel art with on/off62 alpha; the update then keeps that grid and palette).63 `pet_check.py pepper.png --json-out work/before.json` measures it: height per row against idle,64 jump lift, look-row drift. `pet_preview.py contact` and `pet_preview.py gif` show the current65 pet. Ask what should change if the request is vague, naming what the check found.663. **Redo the run.**67 `pet_prepare.py init --name Pepper --id pepper --description "<from pet.json>" --from-split work/split --only running-right,running-left --out work/run`68 (every row of the table starts with an `init --from-split` like this; `--only` names the rows to69 generate). The `match` style keeps the sheet's look; each job's `references` are70 `references/base.png` (the idle frame, enlarged on the key) and `references/rows/<row>.png` (the71 current row). Each job maps onto one `model_run`: its `references` uploaded with `upload_asset`72 and passed as `referenceImages` (an array), its `prompt` verbatim, its `sizes` for the model,73 `background: "opaque"` on Sunburst, priced with `dry_run` first; download the result to the job's74 `output`. When the old row's motion or directions are what is wrong (a jump that never lifts, a75 run that does not alternate, a head that turns the wrong way), add `--no-current-row` to `init`76 so the old strip is not a reference, since the model copies what it is shown, and say what77 changes with `pet_prepare.py note work/run --jobs <row> --text "..."`; a redone look row still78 references the kept other half of the sweep. Then79 `pet_frames.py extract work/run --job running-right`, read the report and preview, and mirror it80 with `--mirror-left` or generate `running-left` the same way.814. **Rebuild.** `pet_build.py work/run --mirror-left`. With the original sheet as the base (recorded82 by `init`), only the rebuilt rows change, scaled to the existing idle row's height, ground line83 and anchor; the edge cleanup touches only them.84 `pet_check.py work/run/final/spritesheet.webp --run work/run` must print `ok: true`; kept look85 rows only bring a warning that they were not reviewed, while a redone look row needs fresh86 verdicts in `qa/direction-semantics.json` (format in the contract) or the check fails.875. **Show and package.**88 `pet_preview.py gif work/run/final/spritesheet.webp --out-dir work/run/previews`, show `pet.gif`,89 upload the sheet and both GIFs into a collection as the create skill does, then90 `pet_package.py make work/run --id pepper`. Ask before `pet_package.py install work/run --force`,91 which backs up the installed Pepper into the run before replacing it.926. **The scarf, later.** A new look changes every frame, so it is a new generation grounded in the93 old pet:94 `init --name Pepper --id pepper --description "<from pet.json>" --from-split work/split --change "add a red knitted scarf" --out work/scarf`95 adds an edit job on the identity image. Approve the edited base, then run every row as96 `scenario-chatgpt-pet-create` does from its rows step (look rows included for a v2 sheet, with97 fresh direction verdicts). The old sheet still sets the pet's size, so the new pet does not grow98 or shrink.99100For a v1 to v2 upgrade (`init --version 2 --only look`), run `look-cardinals`, `look-9` and101`look-10` as the create skill describes and write the direction verdicts; the build keeps rows 0-8102and adds the two look rows.103104## Common mistakes105106- Overwriting the user's file or installed pet: work on a copy, and replace an installed Codex pet107 only after a yes, with `--force` (which backs it up).108- Regenerating the whole pet for one bad row: redo that row; the build copies the others byte for109 byte.110- Sending the whole sheet to an image model as the reference: it reads as a grid of small frames and111 the model copies the grid. Use `references/base.png` and the row strip `split` made.112- Regenerating to fix what the build can fix: leftover pixels, color under transparency and a row113 that drifts are `--clean` and `--reregister`, at no cost.114- Rebuilding without the original sheet: a new row is then sized from nothing; `init --from-split`115 records the sheet so the build keeps the pet's size.116- Changing the pet's id on a rename: Codex finds the pet by its folder, so keep the id and change117 `displayName`.118- Handing a v2 result to someone uploading on ChatGPT web: give them `package/spritesheet-v1.png`.