How c-narcissus/paper-framework-figure-studio-pro works
This repository provides a ChatGPT web-oriented skill for collaboratively producing candidate framework diagrams from a paper. The documented workflow moves from extracting paper facts through strategy, exploratory images, selection, a formal brief, and final candidate images; humans choose and finalize afterward. The README documents these stage relationships, prompt-package audits, and use of first-round images as references in the final round. Skill internals are contained in ZIP files and no source excerpts were provided, so implementation-level wiring is not inferred.
Main entry points: paper-framework-figure-studio-pro-v3.2.15f-skill.zip S0 paper foundation
Components 12
Paper grounding
- Research paper
- S0 paper foundation
paper-framework-figure-studio-pro-v3.2.15f-skill.zip
Exploration
- S1 figure strategy
paper-framework-figure-studio-pro-v3.2.15f-skill.zip - S2 sketch explore
paper-framework-figure-studio-pro-v3.2.15f-skill.zip - Exploratory candidates
- S3 direction select
paper-framework-figure-studio-pro-v3.2.15f-skill.zip
Formal candidates
- S4 candidate brief
paper-framework-figure-studio-pro-v3.2.15f-skill.zip - S5 candidate images
paper-framework-figure-studio-pro-v3.2.15f-skill.zip - Formal candidates
Other components
- Paper author
- Figure Studio skill
paper-framework-figure-studio-pro-v3.2.15f-skill.zip - Human review and finalization
Connections 13
- Paper author → Figure Studio skill: starts workflow
- Paper author → Research paper: provides
- Figure Studio skill → S0 paper foundation: runs
- S0 paper foundation → Research paper: analyzes
- S0 paper foundation → S1 figure strategy: informs
- S1 figure strategy → S2 sketch explore: prepares prompts
- S2 sketch explore → Exploratory candidates: generates C01–C04
- Exploratory candidates → S3 direction select: provides options
- S3 direction select → S4 candidate brief: records preferences
- S4 candidate brief → S5 candidate images: prepares prompts
- Exploratory candidates → S5 candidate images: provides image references
- S5 candidate images → Formal candidates: generates F01–F02
- Formal candidates → Human review and finalization: presents candidates