@1aboveio/skills 0.0.0-stage → 1.0.0

package/skills/content/article-review/references/report-format.md

@@ -0,0 +1,68 @@
+# 评审报告输出格式
+
+使用以下模板输出完整评审报告。
+
+---
+
+```
+# 文章评审报告
+
+**文章标题：** [标题]
+**评审模板：** [模板名称]
+**评审时间：** [日期]
+
+---
+
+## 综合评分：X.X / 10
+（各维度等权平均）
+
+---
+
+## 事实核查结果
+
+- ✅ "[原文陈述]" — 来源：[来源名]
+- ❌ "[原文陈述]" — 实际应为：[正确值]，来源：[来源名]
+- ⚠️ "[原文陈述]" — 无法核实，建议补充来源
+
+---
+
+## 详细评分
+
+### [维度名] X/10
+**依据：**
+- [引用原文 + 分析]
+
+**改进建议：**
+- [具体建议]
+
+[其余维度同上格式...]
+
+---
+
+## 传播潜力分析
+
+**最强传播钩子：**
+> "[从文章中提取或改写的最具传播力的 1-2 句话，可直接用作转发文案]"
+
+**建议传播渠道：**
+- 微信群（行业/专业群）：[适合/不适合，理由一句]
+- 朋友圈：[适合/不适合，理由一句]
+- LinkedIn：[适合/不适合，理由一句]
+
+**传播半径预估：**
+- [ ] 圈内传播（支付/金融科技从业者为主）
+- [ ] 破圈传播（有潜力触达行业外受众）
+
+**提升传播力的快速改动（可选）：**
+- [1-2 条具体改动建议]
+
+---
+
+## 总结
+[2-3 句总体评价，突出最大优势和最需改进的地方]
+
+## 发布建议
+[ ] 可直接发布
+[ ] 修改后发布（需解决上述改进点）
+[ ] 建议重写（存在严重问题）
+```

package/skills/content/article-review/references/template-guide.md

@@ -0,0 +1,79 @@
+# 新建评审模板指南
+
+新建模板时，复制以下结构，保存为 `references/template-<name>.md`，
+然后在 `SKILL.md` 的"可用模板"列表中添加一行。
+
+---
+
+## 模板文件结构
+
+```markdown
+# 评审模板：[模板名称]
+
+**适用范围：** [描述适用的文章类型]
+**目标受众：** [描述文章面向的读者群体]
+**模板版本：** v1.0
+
+---
+
+## 评审维度（Level 1）及细则（Level 2）
+
+### 1. [维度名] [英文名]
+
+[维度说明，1-2句]
+
+**Level 2 细则：**
+- **[细则名]**：[说明]
+- **[细则名]**：[说明]
+- （2-5条）
+
+**评分参考：**
+- 9-10：[优秀标准]
+- 6-8：[良好标准]
+- 3-5：[及格标准]
+- 1-2：[不合格标准]
+
+---
+
+### 2. [下一个维度...]
+```
+
+---
+
+## 设计原则
+
+**维度数量：** 建议 4-6 个 Level 1 维度。维度太多会分散注意力；太少则覆盖不全。
+
+**维度选择：** 每个维度应该：
+- 相互独立（不重叠）
+- 合起来覆盖文章质量的核心方面
+- 与受众和文章类型紧密相关
+
+**事实准确性注意：** 若模板包含"事实准确性"维度，必须包含以下说明：
+> ⚠️ 本维度评分必须基于 web_search 实际核查结果，不接受主观判断。
+
+**Level 2 细则：** 每个维度建议 3-5 条细则，每条清晰说明考察什么。
+
+---
+
+## 模板示例说明
+
+| 场景 | 文件名建议 | 差异化维度 |
+|------|-----------|-----------|
+| 面向 B 端 CFO 的行业报告 | `template-cfo-report.md` | 数据严谨性、决策价值 |
+| 面向年轻用户的品牌故事 | `template-brand-story.md` | 情感温度、品牌一致性 |
+| 合规/政策解读类 | `template-policy.md` | 准确性权重更高、实操指导 |
+| 产品功能介绍 | `template-product.md` | 清晰度、用例覆盖、CTA |
+
+---
+
+## 注册新模板
+
+完成模板文件后，更新 `SKILL.md` 中的"可用模板"列表：
+
+```markdown
+## 可用模板
+
+- **通用** (`references/template-general.md`) — 适用于大多数 [品牌] 公众号内容（默认）
+- **[新模板名]** (`references/template-<name>.md`) — [一句话描述]
+```

package/skills/content/article-review/references/template-wechat-payments.md

@@ -0,0 +1,125 @@
+# 评审模板：公众号-支付行业
+
+**适用范围：** [品牌] 公众号发布的支付/金融科技相关内容文章
+**目标受众：** 支付、金融科技、跨境业务从业者和决策者
+**模板版本：** v1.1
+
+## 评审维度目录
+1. [原创性 Originality](#1-原创性)
+2. [可读性 Readability](#2-可读性)
+3. [叙事与钩子 Narrative Hook](#3-叙事与钩子)
+4. [可传播性 Shareability](#4-可传播性)
+5. [时效性 Timeliness](#5-时效性)
+6. [事实准确性 Factual Accuracy](#6-事实准确性)
+
+---
+
+## 评审维度（Level 1）及细则（Level 2）
+
+### 1. 原创性 Originality
+
+文章是否有独特的观点、角度或内容，而非对已有内容的重复/搬运。
+
+**Level 2 细则：**
+- **观点独特性**：有自己的见解，不是别人文章的复述
+- **角度差异化**：与同类文章相比，切入角度是否新颖
+- **内容增量**：信息或分析是否首发，或相对已有内容有实质增量
+
+**评分参考：**
+- 9-10：完全原创，观点鲜明，与市面内容有明显差异
+- 6-8：有一定原创角度，但部分内容与已有材料相似
+- 3-5：主要是整合/改写，缺乏自有观点
+- 1-2：明显搬运，无实质增量
+
+---
+
+### 2. 可读性 Readability
+
+文章是否易于理解、结构清晰、语言流畅。
+
+**Level 2 细则：**
+- **标题吸引力**：标题是否准确、有吸引力、能反映核心内容
+- **结构清晰度**：段落逻辑是否清晰，有无合理的小标题/分段
+- **语言流畅性**：句子是否通顺，无冗余/歧义
+- **篇幅适当性**：长度是否与内容密度匹配，无注水
+
+**评分参考：**
+- 9-10：标题抓人，结构一目了然，读来流畅，篇幅恰当
+- 6-8：整体可读，有小问题（如段落过长、标题平淡）
+- 3-5：结构混乱或语言拗口，影响阅读体验
+- 1-2：难以阅读，结构缺失或大量错别字/语病
+
+---
+
+### 3. 叙事与钩子 Narrative Hook
+
+文章是否有足够的叙事牵引力，让读者想读完并分享。
+
+**Level 2 细则：**
+- **踩坑叙事密度**：是否有第一人称真实经历作为开场钩子（「我踩了X次坑」「我们处理了N个案例」）；B2B 版可用「客户案例」替代「我」
+- **工具化程度**：踩坑/分析之后，是否提供可操作的步骤、CheckList 或模板，让读者「照着做就行」；纯观点文章无工具化内容得分较低
+- **金句存在性**：是否有至少 1 个可独立截图传播的金句（类比型/反常识型/数字冲击型），放在段首/段尾显眼位置；结尾必须有金句
+
+**评分参考：**
+- 9-10：有明确叙事钩子 + 可操作工具化内容 + 至少 2 个高质量金句
+- 6-8：有钩子或工具化，但不完整（如有步骤但无金句，或有金句但无叙事代入感）
+- 3-5：以陈述为主，缺乏叙事钩子或工具化内容，金句稀少
+- 1-2：纯说教型文章，无叙事、无工具、无金句
+
+---
+
+### 4. 可传播性 Shareability
+
+文章是否具备被读者主动转发分享的动力。
+
+**Level 2 细则：**
+- **情感共鸣**：是否触动读者情绪（共鸣、惊喜、愤慨、感动等）
+- **实用价值**：读者是否能从文章获得可直接使用的知识/工具
+- **社交货币**：分享后能否让读者显得有洞察力、专业或关心行业
+- **话题热度**：是否蹭了有传播势能的话题或事件
+- **受众精准度**：内容与目标读者（支付/金融科技从业者、决策者）的痛点契合程度；越精准，圈内自发传播越强
+- **传播形式适配**：是否有截图友好的金句、结论卡、数据图，利于在微信群/朋友圈二次传播
+- **行动号召清晰度**：是否有明确的 CTA（关注账号、转发、联系我们、扫码），引导读者下一步行动
+
+**评分参考：**
+- 9-10：多维度传播驱动力强——情感共鸣 + 实用价值 + 受众精准 + 传播形式适配，CTA 自然
+- 6-8：有传播价值，但部分维度缺失（如受众精准但缺金句，或有热度但实用性弱）
+- 3-5：主要服务存量读者，缺乏主动转发动力，形式不利于二次传播
+- 1-2：缺乏传播动力，内容封闭，无 CTA，受众定位模糊
+
+---
+
+### 5. 时效性 Timeliness
+
+文章内容是否与当前时间节点相关，发布时机是否恰当。
+
+**Level 2 细则：**
+- **话题新鲜度**：主题是否仍处于公众关注窗口内
+- **发布时机**：相对于事件/节点，发布是领先、同步还是滞后
+- **长尾价值**：若非即时性内容，是否具备长期参考价值
+
+**评分参考：**
+- 9-10：完美踩点，话题正热，发布时机极佳
+- 6-8：时机尚可，话题热度略降或稍超前
+- 3-5：话题冷却期发布，或时机明显滞后
+- 1-2：内容严重过时，或与当下毫无关联
+
+---
+
+### 6. 事实准确性 Factual Accuracy
+
+文章中的陈述、数据和逻辑是否经得起核查。
+
+> ⚠️ 本维度评分必须基于 web_search 实际核查结果，不接受主观判断或"无法核实"的笼统处理。
+
+**Level 2 细则：**
+- **逻辑自洽**：论点与论据之间是否存在矛盾或跳跃
+- **行业知识准确**：对行业概念、术语、常识的描述是否正确
+- **数字真实性**：引用的数据、统计、比例是否有来源，是否与可查资料一致
+- **来源可信度**：引用的信息来源是否权威、可溯源
+
+**评分参考：**
+- 9-10：逻辑严密，所有可查数据均已核实，行业知识准确
+- 6-8：大体准确，有 1-2 处轻微错误或引用不严谨
+- 3-5：多处存疑，数据或逻辑有明显问题
+- 1-2：存在严重事实错误，或核心主张与可查来源矛盾

package/skills/content/article-review/release-check.mjs

@@ -0,0 +1,93 @@
+#!/usr/bin/env node
+import { spawnSync } from 'node:child_process';
+import { mkdirSync, readFileSync, writeFileSync } from 'node:fs';
+import { homedir } from 'node:os';
+import { dirname, join } from 'node:path';
+import { fileURLToPath } from 'node:url';
+
+const DAY_MS = 24 * 60 * 60 * 1000;
+const WEEK_MS = 7 * DAY_MS;
+const here = dirname(fileURLToPath(import.meta.url));
+
+function readJson(path, fallback = {}) {
+  try { return JSON.parse(readFileSync(path, 'utf8')); } catch { return fallback; }
+}
+
+function frontmatterValue(text, key) {
+  const frontmatter = /^---\n([\s\S]*?)\n---/.exec(text)?.[1] ?? '';
+  const match = new RegExp('^' + key + ':\\s*([^\\n]+)$', 'm').exec(frontmatter);
+  return match?.[1].trim().replace(/^['"]|['"]$/g, '') ?? null;
+}
+
+function repositoryOf(source) {
+  const value = String(source ?? '').trim();
+  const match = /^(?:git@github\.com:|https:\/\/github\.com\/)?([A-Za-z0-9_.-]+\/[A-Za-z0-9_.-]+?)(?:\.git)?$/.exec(value);
+  return match?.[1] ?? null;
+}
+
+function stableVersion(value) {
+  const match = /^v?(\d+)\.(\d+)\.(\d+)$/.exec(String(value ?? '').trim());
+  return match ? match.slice(1).map(Number) : null;
+}
+
+function newer(left, right) {
+  for (let index = 0; index < 3; index += 1) {
+    if (left[index] !== right[index]) return left[index] > right[index];
+  }
+  return false;
+}
+
+function writeCache(path, value) {
+  try {
+    mkdirSync(dirname(path), { recursive: true });
+    writeFileSync(path, JSON.stringify(value));
+  } catch { /* A read-only cache must not block skill use. */ }
+}
+
+function main(now = Date.now()) {
+  if (process.env.ONEABOVEIO_NO_UPDATE_CHECK) return;
+  const skill = readFileSync(join(here, 'SKILL.md'), 'utf8');
+  const name = frontmatterValue(skill, 'name');
+  const localVersion = stableVersion(frontmatterValue(skill, 'release-version'));
+  if (!name || !localVersion) return;
+
+  const lock = readJson(join(process.env.HOME || homedir(), '.agents', '.skill-lock.json'));
+  const entry = lock.skills?.[name];
+  const repository = repositoryOf(entry?.sourceUrl || entry?.source);
+  if (!repository) return;
+
+  const cacheRoot = process.env.XDG_CACHE_HOME || join(process.env.HOME || homedir(), '.cache');
+  const cachePath = join(cacheRoot, '1aboveio-skills', 'release-check.json');
+  const all = readJson(cachePath);
+  const cache = all[repository] && typeof all[repository] === 'object' ? all[repository] : {};
+
+  if (!cache.lastCheck || now - cache.lastCheck >= DAY_MS) {
+    const result = spawnSync('gh', ['api', 'repos/' + repository + '/releases/latest', '--jq', '.tag_name'], {
+      encoding: 'utf8',
+      stdio: ['ignore', 'pipe', 'ignore'],
+      timeout: 1500,
+    });
+    cache.lastCheck = now;
+    if (result.status === 0 && stableVersion(result.stdout)) cache.latestTag = result.stdout.trim();
+    else delete cache.latestTag;
+    all[repository] = cache;
+    writeCache(cachePath, all);
+  }
+
+  const latestVersion = stableVersion(cache.latestTag);
+  if (!latestVersion || !newer(latestVersion, localVersion)) return;
+  if (cache.notifiedTag === cache.latestTag && cache.notifiedAt && now - cache.notifiedAt < WEEK_MS) return;
+
+  cache.notifiedTag = cache.latestTag;
+  cache.notifiedAt = now;
+  all[repository] = cache;
+  writeCache(cachePath, all);
+  const command = `npx skills@1.5.23 add '${repository}#${cache.latestTag}' --global --agent universal claude-code --skill ${name} --yes`;
+  process.stdout.write(
+    `UPDATE_AVAILABLE: A newer 1AboveIO skills release (${cache.latestTag}) is available; this skill is `
+      + `v${localVersion.join('.')}. Ask once whether to refresh ${name}. On approval run \`${command}\`. `
+      + 'The refreshed instructions apply next session. Either way, continue the current task and do not ask again this session.\n',
+  );
+}
+
+try { main(); } catch { /* Update checks are advisory and always fail silent. */ }

package/skills/content/article-to-image/LICENSE

@@ -0,0 +1,3 @@
+MIT License
+
+Copyright (c) 2026 1AboveIO

package/skills/content/article-to-image/SKILL.md

@@ -0,0 +1,91 @@
+---
+name: article-to-image
+description: "Generate article cover images: WeChat banner/square (公众号封面), Xiaohongshu (小红书封面), LinkedIn banner, general wide. Brainstorms 4 visual directions from the article, then renders each format with correct aspect ratios and title-safe zones. Use when given an article and asked for covers, banners, or 出图."
+release-version: 0.36.0
+---
+**Updates:** Once/session run `node <skill-base-dir>/release-check.mjs`.
+# Article-to-Image
+
+Generate editorial cover images from article content. Three steps: analyze → brainstorm → generate.
+
+## Format Selection
+
+At the start of every invocation, read `references/formats.md` to get size, aspect ratio, title-safe zone, and prompt suffix for each format.
+
+**Default format**: `wechat-banner` if not specified.
+
+**Multi-format**: If the user requests multiple formats (e.g. "wechat + xiaohongshu"):
+1. Designate the first-mentioned format as the **primary format**
+2. Run the full brainstorm (Step 2) once for the primary format only — do NOT brainstorm per format
+3. After the user selects a direction, adapt that single direction to each additional format (same scene, recomposed for each aspect ratio and title-safe zone)
+
+## Universal Constraints (all formats)
+
+- **Style**: Editorial illustration or cinematic photography — NOT cartoon, anime, or stock photo
+- **Colors**: Max 3 dominant colors per image
+- **No text in image**: Never embed title, captions, or article text
+- **Depth**: Foreground/background separation — avoid flat compositions
+- **Title-safe zone**: Always leave empty space per format spec for text overlay
+
+## Workflow
+
+### Step 1 — Extract core themes from article
+
+Identify:
+- **Central metaphor**: The one concept the article is really about (not the literal topic)
+- **Emotional tone**: Intellectual curiosity / urgency / calm analysis / tension / etc.
+- **Key visual anchors**: 2–3 concrete objects, scenes, or symbols implied by the article
+- **Platform fit**: Does the tone match the requested format's style? (WeChat = serious; XHS = lifestyle)
+
+### Step 2 — Brainstorm 4 visual directions
+
+Each direction must differ in:
+- Primary visual element (scene, object, symbol, environment)
+- Compositional approach (tight portrait / wide environmental / abstract / conceptual)
+- Emotional register
+
+Present as a numbered list before generating. Format:
+
+```
+**Direction 1 — [Name]**
+Visual: [specific scene description]
+Mood: [emotional register]
+Composition: [subject placement + title-safe side]
+Style: [cinematic / editorial illustration / documentary]
+```
+
+Proceed to Step 3 immediately unless user asks to review first.
+
+### Step 3 — Generate images
+
+Build each prompt using this template:
+
+```
+[Specific scene: concrete subject, action, environment],
+[lighting: named style e.g. "dramatic side lighting", "golden hour"],
+[palette: name 2-3 colors],
+[composition: e.g. "subject on right, left third empty for text overlay"],
+[style: editorial photography / cinematic],
+no text, no watermarks, no logos,
+[format prompt suffix from formats.md]
+```
+
+To generate each image, load and follow the `nano-banana-pro` skill, passing the constructed prompt. Generate **one at a time** (not batched) to avoid 503 errors. If a 503/500 error occurs, wait briefly and retry the same image before moving on. Label each result with its direction name.
+
+## Prompt Engineering Rules
+
+**DO:**
+- Use specific concrete scenes: "A taxi driver's hand refusing a banknote, warm amber interior"
+- Name lighting explicitly: "golden hour side light", "cool blue overhead", "dramatic chiaroscuro"
+- Reference photographic styles: "editorial photography", "documentary", "cinematic wide shot"
+- Specify negative space: "empty left third for title overlay"
+
+**DON'T:**
+- Translate article topic literally into image (伊朗支付 ≠ 伊朗国旗 + 信用卡)
+- Use vague adjectives alone: "beautiful", "professional", "modern"
+- Include any text or UI elements
+- Use styles that clash with the format's tone
+
+## Format Reference
+
+See `references/formats.md` for all supported formats with sizes, title-safe zones, and prompt suffixes.

package/skills/content/article-to-image/references/formats.md

@@ -0,0 +1,53 @@
+# Output Format Specifications
+
+## Supported Formats
+
+### `wechat-banner` — 微信公众号首图
+- **Size**: 900×383px (2.35:1)
+- **Use case**: 文章顶部大图，分享卡片缩略图
+- **Title-safe zone**: Left third empty OR top quarter empty
+- **Style notes**: Editorial / cinematic, serious tone, industry-appropriate
+- **Prompt suffix**: `wide format 2.35:1, professional WeChat article banner, editorial photography`
+
+### `wechat-square` — 微信公众号方形封面
+- **Size**: 900×900px (1:1)
+- **Use case**: 公众号封面缩略图（列表页显示）
+- **Title-safe zone**: Bottom quarter for title overlay
+- **Style notes**: Strong single focal point, works at small size
+- **Prompt suffix**: `square format 1:1, WeChat thumbnail cover, bold focal point`
+
+### `xiaohongshu-cover` — 小红书封面
+- **Size**: 1080×1350px (4:5)
+- **Use case**: 小红书笔记封面图
+- **Title-safe zone**: Top 20% and bottom 20% for text overlay
+- **Style notes**: More lifestyle/aesthetic, warmer palette acceptable, slightly more vibrant than WeChat
+- **Prompt suffix**: `portrait format 4:5, Xiaohongshu cover image, lifestyle editorial, vibrant but refined`
+
+### `xiaohongshu-square` — 小红书方形
+- **Size**: 1080×1080px (1:1)
+- **Use case**: 小红书方形笔记封面
+- **Title-safe zone**: Bottom third
+- **Style notes**: Same as xiaohongshu-cover
+- **Prompt suffix**: `square format 1:1, Xiaohongshu post cover, lifestyle editorial`
+
+### `linkedin-banner` — LinkedIn 文章封面
+- **Size**: 1200×627px (1.91:1)
+- **Use case**: LinkedIn article header, link preview
+- **Title-safe zone**: Center or left half
+- **Style notes**: Corporate but not sterile, professional editorial
+- **Prompt suffix**: `wide format 1.91:1, LinkedIn article banner, professional editorial photography`
+
+### `general-wide` — 通用宽幅
+- **Size**: 1920×1080px (16:9)
+- **Use case**: Presentation slides, multi-platform reuse
+- **Title-safe zone**: Left third or top quarter
+- **Style notes**: Cinematic, high-impact
+- **Prompt suffix**: `widescreen 16:9 format, cinematic, versatile editorial banner`
+
+## Default
+
+If no format is specified, default to `wechat-banner`.
+
+## Adding New Formats
+
+To add a new format, append a new section following the same structure: Size, Use case, Title-safe zone, Style notes, Prompt suffix.

package/skills/content/article-to-image/release-check.mjs

@@ -0,0 +1,93 @@
+#!/usr/bin/env node
+import { spawnSync } from 'node:child_process';
+import { mkdirSync, readFileSync, writeFileSync } from 'node:fs';
+import { homedir } from 'node:os';
+import { dirname, join } from 'node:path';
+import { fileURLToPath } from 'node:url';
+
+const DAY_MS = 24 * 60 * 60 * 1000;
+const WEEK_MS = 7 * DAY_MS;
+const here = dirname(fileURLToPath(import.meta.url));
+
+function readJson(path, fallback = {}) {
+  try { return JSON.parse(readFileSync(path, 'utf8')); } catch { return fallback; }
+}
+
+function frontmatterValue(text, key) {
+  const frontmatter = /^---\n([\s\S]*?)\n---/.exec(text)?.[1] ?? '';
+  const match = new RegExp('^' + key + ':\\s*([^\\n]+)$', 'm').exec(frontmatter);
+  return match?.[1].trim().replace(/^['"]|['"]$/g, '') ?? null;
+}
+
+function repositoryOf(source) {
+  const value = String(source ?? '').trim();
+  const match = /^(?:git@github\.com:|https:\/\/github\.com\/)?([A-Za-z0-9_.-]+\/[A-Za-z0-9_.-]+?)(?:\.git)?$/.exec(value);
+  return match?.[1] ?? null;
+}
+
+function stableVersion(value) {
+  const match = /^v?(\d+)\.(\d+)\.(\d+)$/.exec(String(value ?? '').trim());
+  return match ? match.slice(1).map(Number) : null;
+}
+
+function newer(left, right) {
+  for (let index = 0; index < 3; index += 1) {
+    if (left[index] !== right[index]) return left[index] > right[index];
+  }
+  return false;
+}
+
+function writeCache(path, value) {
+  try {
+    mkdirSync(dirname(path), { recursive: true });
+    writeFileSync(path, JSON.stringify(value));
+  } catch { /* A read-only cache must not block skill use. */ }
+}
+
+function main(now = Date.now()) {
+  if (process.env.ONEABOVEIO_NO_UPDATE_CHECK) return;
+  const skill = readFileSync(join(here, 'SKILL.md'), 'utf8');
+  const name = frontmatterValue(skill, 'name');
+  const localVersion = stableVersion(frontmatterValue(skill, 'release-version'));
+  if (!name || !localVersion) return;
+
+  const lock = readJson(join(process.env.HOME || homedir(), '.agents', '.skill-lock.json'));
+  const entry = lock.skills?.[name];
+  const repository = repositoryOf(entry?.sourceUrl || entry?.source);
+  if (!repository) return;
+
+  const cacheRoot = process.env.XDG_CACHE_HOME || join(process.env.HOME || homedir(), '.cache');
+  const cachePath = join(cacheRoot, '1aboveio-skills', 'release-check.json');
+  const all = readJson(cachePath);
+  const cache = all[repository] && typeof all[repository] === 'object' ? all[repository] : {};
+
+  if (!cache.lastCheck || now - cache.lastCheck >= DAY_MS) {
+    const result = spawnSync('gh', ['api', 'repos/' + repository + '/releases/latest', '--jq', '.tag_name'], {
+      encoding: 'utf8',
+      stdio: ['ignore', 'pipe', 'ignore'],
+      timeout: 1500,
+    });
+    cache.lastCheck = now;
+    if (result.status === 0 && stableVersion(result.stdout)) cache.latestTag = result.stdout.trim();
+    else delete cache.latestTag;
+    all[repository] = cache;
+    writeCache(cachePath, all);
+  }
+
+  const latestVersion = stableVersion(cache.latestTag);
+  if (!latestVersion || !newer(latestVersion, localVersion)) return;
+  if (cache.notifiedTag === cache.latestTag && cache.notifiedAt && now - cache.notifiedAt < WEEK_MS) return;
+
+  cache.notifiedTag = cache.latestTag;
+  cache.notifiedAt = now;
+  all[repository] = cache;
+  writeCache(cachePath, all);
+  const command = `npx skills@1.5.23 add '${repository}#${cache.latestTag}' --global --agent universal claude-code --skill ${name} --yes`;
+  process.stdout.write(
+    `UPDATE_AVAILABLE: A newer 1AboveIO skills release (${cache.latestTag}) is available; this skill is `
+      + `v${localVersion.join('.')}. Ask once whether to refresh ${name}. On approval run \`${command}\`. `
+      + 'The refreshed instructions apply next session. Either way, continue the current task and do not ask again this session.\n',
+  );
+}
+
+try { main(); } catch { /* Update checks are advisory and always fail silent. */ }

package/skills/content/flowchart-svg/LICENSE

@@ -0,0 +1,3 @@
+MIT License
+
+Copyright (c) 2026 1AboveIO

package/skills/content/flowchart-svg/SKILL.md

@@ -0,0 +1,43 @@
+---
+name: flowchart-svg
+description: Create polished flowcharts directly as SVG and export PNG copies. Use when asked to draw, design, render, or deliver process diagrams, architecture diagrams, consulting workflows, decision flows, swimlanes, or other flowcharts as image files, especially when final outputs must include .svg and .png versions.
+release-version: 0.36.0
+---
+**Updates:** Once/session run `node <skill-base-dir>/release-check.mjs`.
+# Flowchart SVG
+
+Use this skill to produce clean, editable SVG flowcharts and matching PNG exports.
+
+## Workflow
+
+1. Identify the diagram purpose, audience, title, nodes, edges, and feedback loops.
+2. Choose a simple layout before writing SVG: left-to-right for pipelines, top-to-bottom for procedures.
+3. Draw directly in SVG with explicit coordinates; avoid Mermaid screenshots unless requested.
+4. Use semantic color groups: input, AI/automation, human judgment, reusable assets, outputs, feedback.
+5. Keep text short; split long labels into multiple text lines rather than relying on wrapping.
+6. Export PNG from the SVG with rsvg-convert or the bundled shell helper.
+7. Send the PNG to the user by default, because SVG is often unsupported in chat and document surfaces.
+8. Include the SVG as an editable source file when the channel supports attachments or when the user explicitly asks for it.
+9. Inspect both outputs before delivery.
+
+## Quality Bar
+
+- Include a title and one-line subtitle when helpful.
+- Use viewBox, markers for arrows, rounded rectangles, and consistent spacing.
+- Use readable fonts with fallbacks, normally Arial, Noto Sans CJK SC, sans-serif.
+- Avoid tiny text; body text should normally be at least 13px in a 1200px-wide diagram.
+- Keep arrows from crossing boxes where possible.
+
+## Export
+
+Use the helper script:
+
+```bash
+scripts/export_png.sh input.svg output.png
+```
+
+If the helper is unavailable, run:
+
+```bash
+rsvg-convert -w 2400 -f png input.svg -o output.png
+```