← 提示词库 Meta/muse-agent/skills/artifacts/pdf/references/visual.md 原文 md
🌐 中英双语对照

description: Resolving user-provided visual direction into the style plan for PDF artifacts.

PDF Visual Guidance / PDF 视觉指南

User-provided visual direction in verbatim_request or the parent conversation
is authoritative. Before styling, take one pass against generic defaults:
make each visual choice because it fits this deliverable, not because it
is the easiest template (one accent everywhere, a single font doing all
the work, no hierarchy between title and body, emoji as icons).

verbatim_request 或父对话中用户提供的视觉方向具有权威性。在设置样式之前,先对照通用默认做法过一遍:每个视觉选择都应该因为它适合这个交付物而做出,而不是因为它是最省事的模板(到处用同一个强调色、让单一字体承担一切、标题与正文没有层级、把 emoji 当图标)。

If a delivered PDF has a reported visual problem, re-render the affected pages
to PNG and read them before changing the source. Do not guess at layout or
styling defects without visual evidence.

如果交付的 PDF 被报告存在视觉问题,先把受影响的页面重新渲染为 PNG 并查看,然后再修改源文件。没有视觉证据时不要猜测排版或样式缺陷。