Skip to content

AgentFigureGallery v0.1.0

Latest
Latest

Choose a tag to compare

View all tags
@Dsadd4 Dsadd4 released this 01 May 17:14
· 8 commits to main since this release
v0.1.0
b268a30

AgentFigureGallery v0.1.0

AgentFigureGallery is a human-preference reference memory for scientific plotting agents.

It lets Codex, Claude Code, Cursor, and other coding agents query real Nature/Cell/Science-style visual references, show candidates to a human, record like/reject/select feedback, and export a selected bundle before plotting code is written.

Install

One-command Codex bootstrap:

curl -fsSL https://raw.githubusercontent.com/Dsadd4/AgentFigureGallery/main/scripts/install.sh | bash

Install all agent entrypoints:

curl -fsSL https://raw.githubusercontent.com/Dsadd4/AgentFigureGallery/main/scripts/install.sh | AFG_AGENT_TARGETS="codex claude-code cursor" bash

Cursor Project Rule:

agentfiguregallery install-cursor-rule --project /path/to/your-cursor-project

Full public KB:

agentfiguregallery setup --pack full-public --manifest-url https://huggingface.co/datasets/dsadd4/AgentFigureGallery/resolve/main/resource_manifest.json

Highlights

  • One-command bootstrap with Python 3.10+ auto-detection.
  • Codex skill install at ~/.codex/skills/agent-figure-gallery.
  • Claude Code skill install at ~/.claude/skills/agent-figure-gallery.
  • Cursor-compatible skill install and official Cursor Project Rule output at .cursor/rules/agent-figure-gallery.mdc.
  • Dynamic browser gallery for candidate generation, global/type-level preference memory, and bundle export.
  • 16,341 public scientific figure references across 10 plot types.
  • Hugging Face full-public KB distribution with GitHub Release fallback manifest.
  • Before/after benchmark showing prompt-only plotting vs reference-guided plotting.

Visual Demo

AgentFigureGallery dynamic demo

Before after benchmark

Agent Workflow

agent query -> gallery display -> human like/reject/select -> export bundle -> agent action

Recommended agent prompt:

Read the agent-figure-gallery skill, then use AgentFigureGallery before writing publication figure code.

Verification

The release commit passes:

  • Python compile
  • private-path scan
  • Codex / Claude Code / Cursor-compatible install smoke tests
  • Cursor Project Rule generation
  • query / gallery / prefer / bundle smoke test
  • Hugging Face dataset card sync dry-run

Links

Assets 2
Loading