Menu

Individual skill

stickerforge

0upvotes

Turns one character image into hundreds of print-ready transparent PNG stickers. Plans concepts, checks duplicates, runs a pilot, batch-generates with Codex image generation or an OpenAI/Gemini key, gates alpha and fake-checkerboard backgrounds, retries failures twice, exports a contact sheet and die-cut notes. Use for sticker packs, character sticker sets, emoji or messenger stickers, print stickers, mascot sticker sheets. Korean triggers - 스티커 만들어, 스티커 팩, 캐릭터 스티커 대량 생성, 이모티콘 스티커, 인쇄용 스티커, 칼선, 스티커 시트, 투명 PNG 스티커, 컨셉 100개.

skills/stickerforge/SKILL.md

Read pinned source

Source commit 7f4ef91840a66b72de444b5c25fb5f14c97218a8

What was checked

Matches the observed repository HEAD
Yes
Scanner rules are current
Yes
Repository scan complete
Yes
Skill scan complete
Yes
Observed behavior
Network or broad shell
Last checked
Nov 11, 2026

These are observations at the recorded commit, not a guarantee of safety. Review the source before running it.

Install this skill

Choose an agent in your Stack or copy a command below. Commands fetch the pinned commit and copy only this skill directory. Existing destinations are refused. Copying a command does not install anything.

An install command is unavailable: current rules, complete repository and skill scans, an eligible behavior level and a permitted license are required.

Repository context

Repository stars (not skill ratings)
37
Repository last updated
Nov 6, 2026
License
MIT

Observed agent integration: No specific integration evidence. Standard SKILL.md installation does not imply verified compatibility.

Files and excerpt

File metadata and permitted, redacted excerpts come from the recorded scan. Source links are pinned to that commit.

Browse files
Read excerpt

# stickerforge

One reference character in, a gated set of transparent PNG stickers out. Work in a project folder outside this skill folder. Scripts live in `scripts/` and are run, not read. Run `scripts/doctor.py` first; it detects the runtime and stops when image generation is unavailable.

## Invariants

1. Three human checks only: style (A), pilot (B), final (C). No batch run before B.
2. The reference is always the original character image. Generated images are never used as references.
3. Transparent output: ask for `transparent_background=true`, then clean alpha (12 and below to 0, 240 and above to 255). A checkerboard painted into pixels is a fail, not a transparent image.
4. Never add text on top of an image afterwards. A wrong glyph means regenerate.
5. Concepts are compared against the full concept list before generation. Duplicates are dropped.
6. Front model for the Codex path: gpt-6.1-terra at medium effort. Pass the prompt verbatim. No lighter model; quality drops in practice.
7. Collect images by the `thread_id` from `codex exec --json`, never by time window.
8. Resume from `manifest.jsonl`: append one row per state change, the last row for an id wins. Retry limit is 2 per sticker.
Back to top