concept-anchor-figure
Skill0 starsChanged 2 months agoMITView on GitHub
Turn one article's core idea into a single watermark-free vector infographic (HTML -> PNG via fig_fit), in the workspace brand palette (white bg, black/red/gray + journey blue #056DE8). Use when the user asks for a concept anchor figure, "one picture to understand this article", a sketchnote-style summary figure, an abstract illustration with crisp Chinese text, or a watermark-free illustration instead of AI-generated images. Complements pub-infographic-supplement (density backfill); this skill makes ONE anchor figure that explains the article, not per-section density.
In this repository139 other files it gives its agents.
- AGENTS.mdAGENTS.md
- skills/caveman-learn/CLAUDE.mdCLAUDE.md
- adversarial-reviewSkill
- ai-repo-engineering-layoutSkill
- ai-session-material-minerSkill
- ai-usage-reportSkill
- ardot-effect-fallbackSkill
- audio-friendly-adaptSkill
- automation-resurrectionSkill
- book2vido-videoSkill
- boss-autopilotSkill
- brand-icon-sourcingSkill
- canvas-designSkill
- cavecrewSkill
- caveman-commitSkill
- caveman-compressSkill
- caveman-discoverSkill
- caveman-evidence-reviewSkill
- caveman-exploreSkill
- caveman-helpSkill
- caveman-learnSkill
- caveman-manageSkill
- caveman-optimizeSkill
- caveman-reviewSkill
- caveman-setupSkill
- cavemanSkill
- caveman-statsSkill
- cc-switch-usage-auditSkill
- certification-submission-kitSkill
- comment-reply-playbookSkill
- conference-archiveSkill
- constitution-amendmentSkill
- content-format-auditSkill
- content-matrix-channel-auditSkill
- cross-repo-skill-curationSkill
- daily-judgment-briefSkill
- daily-reportSkill
- de-ai-flavorSkill
- desktop-ai-forensicsSkill
- discernment-nudgeSkill
- doc-translateSkill
- doubao-app-layer-archiveSkill
- effort-accountingSkill
- ego-browserSkill
- egress-allowlistSkill
- eval-skillsSkill
- evolve-skillsSkill
- exam-prep-pmpSkill
- fact-check-gateSkill
- fanout-mainline-ledgerSkill
- free-token-channel-auditSkill
- github-repo-gallerySkill
- git-identity-pinSkill
- git-push-in-sandboxSkill
- harness-governanceSkill
- harness-model-upgrade-auditSkill
- hidden-contract-auditSkill
- investigate-firstSkill
- ip-platform-fit-auditSkill
- lean-buildSkill
- The first 60 shown.
The file
Files
概念锚点图(concept-anchor-figure)
适用守卫:专为「行途 XingTu」工作区设计(fig_fit/发布包结构/品牌色)。先读根 AGENTS.md 含「行途 XingTu」才继续;其他空间声明不适用。
何时用 / 何时不用
- 用:一篇文章要一张"一眼看懂"的总结图;概念科普类(方法论/框架/分层/流程隐喻);用户明确要无水印、中文清晰的插图。
- 不用:逐节补图文密度(走
pub-infographic-supplement);头像/封面氛围图(可走生成模型,但注意其水印与中文糊字风险);带精确数据的图表(数据图也走 HTML,但属密度补齐场景)。
设计定则(来自 boss 2026-09-16 拍板)
- 零水印:纯 HTML/CSS 渲染,非生成模型;天然无水印,可随时改字改色重渲。
- 概念图少文字:文字交给正文,图只放"标签级"短语(每卡 ≤2 行);不放数据表(那是密度补齐的活)。
- 品牌色锁死:白底 #FFFFFF、黑 #141414、红 #d71a1b、行途蓝 #056DE8、灰 #555B62/A8ADB3(对齐 DESIGN.md 黑红灰体系 + P028)。
- 一张图一个隐喻:从正文提炼一个可转述的核心结构(四道关卡/传送带/漏斗/三层塔…),不贪多。
五步流程
- 读正文提炼隐喻:通读正文(发布包
01_正文/正文_v<最大>.md),找出"这篇文章用一句话说是什么结构"(反常识点优先)。选不出清晰隐喻 → 停下报告,不硬凑。 - 选模板或自建:
templates/template_gate-flow.html(关卡/漏斗/流程隐喻,四卡横排+贯穿箭头+起终点)可直接改;其他隐喻照其结构自建,保持同一 CSS 骨架(头部一行式标题+右侧灰 tag / 主体 / 底部 motto+署名)。 - 填内容(标签级):每卡只放「名称 + 一句症状 + 一句钥匙」;标题红字点题;底部收三词 motto;右下署名
行途 XingTu · <合集名>。 - 渲染:
cd ${WORKSPACE} && python3 tools/fig_fit.py "<html路径>" --width 1280 --bg "#FFFFFF"(必须从工作区根跑相对路径,实测进包内目录跑会路径解析失败)。产出同名 .png。 - 亲眼看图验收:Read 打开 PNG 确认:无水印/无乱码/箭头与卡片不错位/中文清晰/品牌色正确——不许不看图就交付。通过后 PNG 落发布包
04_配图/,命名<两位序号>-概念锚点_<主题>.png(接续现有最大序号);需要入稿则按 md2wechat 占位规则同步插 MD 与排版 HTML(参照 pub-infographic-supplement 的 before/after 锚点法)。
模板速改指引(gate-flow)
- 4 卡:改
.gate .no(关卡 n)、.nm(名称)、.sym(症状一句)、.key(钥匙一句);少于/多于 4 关改.gates的grid-template-columns。 - 起终点:改
.start .dot(左起点)与.finish .lb(右终点);箭头自动贯穿。 - 标题:
<h1>主标题 +<span class="r">红字强调;右侧.tag写图类型标签。 - 隐喻变体:漏斗=卡宽递减;传送带=起点改"输入"、终点改"交付";塔=竖排 grid。
红线
- 🔴 文字精确一律 HTML 覆字,不用 AI 生图(boss 10-12 ADV-02 沉淀):ADV-02 三版 AI 封面候选 OCR 抓出「杠精→红娘」「红哨错字」三次文字硬伤。本技能本来就是 HTML→PNG 管线(天然符合),但要强化:图上任何文字(标题/标签/数字/引语)都在 HTML 里用 CSS 精确写,再 fig_fit 渲;AI 生图只可做简笔画人物/场景构图参考,绝不承担最终文字。与
wechat-cover-genv1.1.0 铁律 1/2、pub-infographic-supplementv1.2.3 引子图段同口径。 - 🔴 若本图被选为引子图/封面底图,必须与封面同步(坑位 67/72):定稿后
cp <本图>.png → 04_配图/00-封面引子图.png,并确认03_封面/封面.png视觉同源;改封面必同步本图。 - 禁用生成模型出最终稿(水印/糊中文风险);生成图只允许做风格参考。
- 不放公司/项目真名与内部数据(与正文同一脱敏标准,过
privacy_scan.py口径)。 - 每张图交付前必须 Read 自看(视觉验收不可跳过)。
收尾留痕
python3 tools/changelog_append.py --type "🎨" --path "<产物png路径>" --desc "概念锚点图:<文章>·<隐喻>·fig_fit 无水印渲染" --tool qwenwork(或按所在工作区惯例)。
<!-- public-sync: 2026-09-27, 2026-10-12 | 脱敏版本 | 源 .agents/skills/concept-anchor-figure(10-12 增量:文字 HTML 覆字铁律 + 引子图同步铁律) -->Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

