@1aboveio/skills 0.20.4 → 0.21.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +10 -4
- package/package.json +1 -1
- package/runtime/skills/distribution/generated/recipes.json +136 -24
- package/runtime/skills/distribution/scripts/bundles.mjs +56 -27
- package/runtime/skills/engineering/engineering-runtime/scripts/workflow-policy.mjs +1 -1
- package/skills/compliance/doc-authenticity/SKILL.md +2 -1
- package/skills/compliance/doc-authenticity/agents/openai.yaml +8 -0
- package/skills/content-creator/article-review/LICENSE +3 -0
- package/skills/content-creator/article-review/SKILL.md +65 -0
- package/skills/content-creator/article-review/references/report-format.md +68 -0
- package/skills/content-creator/article-review/references/template-guide.md +79 -0
- package/skills/content-creator/article-review/references/template-wechat-payments.md +125 -0
- package/skills/content-creator/article-to-image/LICENSE +3 -0
- package/skills/content-creator/article-to-image/SKILL.md +90 -0
- package/skills/content-creator/article-to-image/references/formats.md +53 -0
- package/skills/content-creator/wechat-article-writer/LICENSE +3 -0
- package/skills/content-creator/wechat-article-writer/SKILL.md +50 -0
- package/skills/content-creator/wechat-article-writer/references/style-comparison.md +64 -0
- package/skills/content-creator/wechat-article-writer/references/style-howto.md +64 -0
- package/skills/content-creator/wechat-article-writer/references/style-insight.md +63 -0
- package/skills/content-creator/wechat-article-writer/references/style-pitfall.md +60 -0
- package/skills/content-creator/wechat-article-writer/references/style-selector.md +133 -0
- package/skills/content-creator/wechat-article-writer/references/writing-rules.md +150 -0
- package/skills/content-creator/wechat-publisher/LICENSE +3 -0
- package/skills/content-creator/wechat-publisher/SKILL.md +170 -0
- package/skills/content-creator/wechat-publisher/references/api.md +105 -0
- package/skills/content-creator/wechat-publisher/references/layout-rules.md +57 -0
- package/skills/content-creator/wechat-publisher/references/wechat-safe-template.html +86 -0
- package/skills/content-creator/wechat-publisher/scripts/md_to_wechat_html.py +194 -0
- package/skills/content-creator/wechat-publisher/scripts/publish.py +239 -0
- package/skills/content-creator/wechat-search/LICENSE +3 -0
- package/skills/content-creator/wechat-search/SKILL.md +210 -0
- package/skills/content-creator/wechat-search/scripts/start-chrome.sh +83 -0
- package/skills/engineering/engineering-runtime/coherence/workflow.json +13 -13
- package/skills/engineering/engineering-runtime/scripts/workflow-policy.mjs +1 -1
|
@@ -0,0 +1,68 @@
|
|
|
1
|
+
# 评审报告输出格式
|
|
2
|
+
|
|
3
|
+
使用以下模板输出完整评审报告。
|
|
4
|
+
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
```
|
|
8
|
+
# 文章评审报告
|
|
9
|
+
|
|
10
|
+
**文章标题:** [标题]
|
|
11
|
+
**评审模板:** [模板名称]
|
|
12
|
+
**评审时间:** [日期]
|
|
13
|
+
|
|
14
|
+
---
|
|
15
|
+
|
|
16
|
+
## 综合评分:X.X / 10
|
|
17
|
+
(各维度等权平均)
|
|
18
|
+
|
|
19
|
+
---
|
|
20
|
+
|
|
21
|
+
## 事实核查结果
|
|
22
|
+
|
|
23
|
+
- ✅ "[原文陈述]" — 来源:[来源名]
|
|
24
|
+
- ❌ "[原文陈述]" — 实际应为:[正确值],来源:[来源名]
|
|
25
|
+
- ⚠️ "[原文陈述]" — 无法核实,建议补充来源
|
|
26
|
+
|
|
27
|
+
---
|
|
28
|
+
|
|
29
|
+
## 详细评分
|
|
30
|
+
|
|
31
|
+
### [维度名] X/10
|
|
32
|
+
**依据:**
|
|
33
|
+
- [引用原文 + 分析]
|
|
34
|
+
|
|
35
|
+
**改进建议:**
|
|
36
|
+
- [具体建议]
|
|
37
|
+
|
|
38
|
+
[其余维度同上格式...]
|
|
39
|
+
|
|
40
|
+
---
|
|
41
|
+
|
|
42
|
+
## 传播潜力分析
|
|
43
|
+
|
|
44
|
+
**最强传播钩子:**
|
|
45
|
+
> "[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]"
|
|
46
|
+
|
|
47
|
+
**建议传播渠道:**
|
|
48
|
+
- 微信群(行业/专业群):[适合/不适合,理由一句]
|
|
49
|
+
- 朋友圈:[适合/不适合,理由一句]
|
|
50
|
+
- LinkedIn:[适合/不适合,理由一句]
|
|
51
|
+
|
|
52
|
+
**传播半径预估:**
|
|
53
|
+
- [ ] 圈内传播(支付/金融科技从业者为主)
|
|
54
|
+
- [ ] 破圈传播(有潜力触达行业外受众)
|
|
55
|
+
|
|
56
|
+
**提升传播力的快速改动(可选):**
|
|
57
|
+
- [1-2 条具体改动建议]
|
|
58
|
+
|
|
59
|
+
---
|
|
60
|
+
|
|
61
|
+
## 总结
|
|
62
|
+
[2-3 句总体评价,突出最大优势和最需改进的地方]
|
|
63
|
+
|
|
64
|
+
## 发布建议
|
|
65
|
+
[ ] 可直接发布
|
|
66
|
+
[ ] 修改后发布(需解决上述改进点)
|
|
67
|
+
[ ] 建议重写(存在严重问题)
|
|
68
|
+
```
|
|
@@ -0,0 +1,79 @@
|
|
|
1
|
+
# 新建评审模板指南
|
|
2
|
+
|
|
3
|
+
新建模板时,复制以下结构,保存为 `references/template-<name>.md`,
|
|
4
|
+
然后在 `SKILL.md` 的"可用模板"列表中添加一行。
|
|
5
|
+
|
|
6
|
+
---
|
|
7
|
+
|
|
8
|
+
## 模板文件结构
|
|
9
|
+
|
|
10
|
+
```markdown
|
|
11
|
+
# 评审模板:[模板名称]
|
|
12
|
+
|
|
13
|
+
**适用范围:** [描述适用的文章类型]
|
|
14
|
+
**目标受众:** [描述文章面向的读者群体]
|
|
15
|
+
**模板版本:** v1.0
|
|
16
|
+
|
|
17
|
+
---
|
|
18
|
+
|
|
19
|
+
## 评审维度(Level 1)及细则(Level 2)
|
|
20
|
+
|
|
21
|
+
### 1. [维度名] [英文名]
|
|
22
|
+
|
|
23
|
+
[维度说明,1-2句]
|
|
24
|
+
|
|
25
|
+
**Level 2 细则:**
|
|
26
|
+
- **[细则名]**:[说明]
|
|
27
|
+
- **[细则名]**:[说明]
|
|
28
|
+
- (2-5条)
|
|
29
|
+
|
|
30
|
+
**评分参考:**
|
|
31
|
+
- 9-10:[优秀标准]
|
|
32
|
+
- 6-8:[良好标准]
|
|
33
|
+
- 3-5:[及格标准]
|
|
34
|
+
- 1-2:[不合格标准]
|
|
35
|
+
|
|
36
|
+
---
|
|
37
|
+
|
|
38
|
+
### 2. [下一个维度...]
|
|
39
|
+
```
|
|
40
|
+
|
|
41
|
+
---
|
|
42
|
+
|
|
43
|
+
## 设计原则
|
|
44
|
+
|
|
45
|
+
**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。
|
|
46
|
+
|
|
47
|
+
**维度选择:** 每个维度应该:
|
|
48
|
+
- 相互独立(不重叠)
|
|
49
|
+
- 合起来覆盖文章质量的核心方面
|
|
50
|
+
- 与受众和文章类型紧密相关
|
|
51
|
+
|
|
52
|
+
**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:
|
|
53
|
+
> ⚠️ 本维度评分必须基于 web_search 实际核查结果,不接受主观判断。
|
|
54
|
+
|
|
55
|
+
**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。
|
|
56
|
+
|
|
57
|
+
---
|
|
58
|
+
|
|
59
|
+
## 模板示例说明
|
|
60
|
+
|
|
61
|
+
| 场景 | 文件名建议 | 差异化维度 |
|
|
62
|
+
|------|-----------|-----------|
|
|
63
|
+
| 面向 B 端 CFO 的行业报告 | `template-cfo-report.md` | 数据严谨性、决策价值 |
|
|
64
|
+
| 面向年轻用户的品牌故事 | `template-brand-story.md` | 情感温度、品牌一致性 |
|
|
65
|
+
| 合规/政策解读类 | `template-policy.md` | 准确性权重更高、实操指导 |
|
|
66
|
+
| 产品功能介绍 | `template-product.md` | 清晰度、用例覆盖、CTA |
|
|
67
|
+
|
|
68
|
+
---
|
|
69
|
+
|
|
70
|
+
## 注册新模板
|
|
71
|
+
|
|
72
|
+
完成模板文件后,更新 `SKILL.md` 中的"可用模板"列表:
|
|
73
|
+
|
|
74
|
+
```markdown
|
|
75
|
+
## 可用模板
|
|
76
|
+
|
|
77
|
+
- **通用** (`references/template-general.md`) — 适用于大多数 [品牌] 公众号内容(默认)
|
|
78
|
+
- **[新模板名]** (`references/template-<name>.md`) — [一句话描述]
|
|
79
|
+
```
|
|
@@ -0,0 +1,125 @@
|
|
|
1
|
+
# 评审模板:公众号-支付行业
|
|
2
|
+
|
|
3
|
+
**适用范围:** [品牌] 公众号发布的支付/金融科技相关内容文章
|
|
4
|
+
**目标受众:** 支付、金融科技、跨境业务从业者和决策者
|
|
5
|
+
**模板版本:** v1.1
|
|
6
|
+
|
|
7
|
+
## 评审维度目录
|
|
8
|
+
1. [原创性 Originality](#1-原创性)
|
|
9
|
+
2. [可读性 Readability](#2-可读性)
|
|
10
|
+
3. [叙事与钩子 Narrative Hook](#3-叙事与钩子)
|
|
11
|
+
4. [可传播性 Shareability](#4-可传播性)
|
|
12
|
+
5. [时效性 Timeliness](#5-时效性)
|
|
13
|
+
6. [事实准确性 Factual Accuracy](#6-事实准确性)
|
|
14
|
+
|
|
15
|
+
---
|
|
16
|
+
|
|
17
|
+
## 评审维度(Level 1)及细则(Level 2)
|
|
18
|
+
|
|
19
|
+
### 1. 原创性 Originality
|
|
20
|
+
|
|
21
|
+
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。
|
|
22
|
+
|
|
23
|
+
**Level 2 细则:**
|
|
24
|
+
- **观点独特性**:有自己的见解,不是别人文章的复述
|
|
25
|
+
- **角度差异化**:与同类文章相比,切入角度是否新颖
|
|
26
|
+
- **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
|
|
27
|
+
|
|
28
|
+
**评分参考:**
|
|
29
|
+
- 9-10:完全原创,观点鲜明,与市面内容有明显差异
|
|
30
|
+
- 6-8:有一定原创角度,但部分内容与已有材料相似
|
|
31
|
+
- 3-5:主要是整合/改写,缺乏自有观点
|
|
32
|
+
- 1-2:明显搬运,无实质增量
|
|
33
|
+
|
|
34
|
+
---
|
|
35
|
+
|
|
36
|
+
### 2. 可读性 Readability
|
|
37
|
+
|
|
38
|
+
文章是否易于理解、结构清晰、语言流畅。
|
|
39
|
+
|
|
40
|
+
**Level 2 细则:**
|
|
41
|
+
- **标题吸引力**:标题是否准确、有吸引力、能反映核心内容
|
|
42
|
+
- **结构清晰度**:段落逻辑是否清晰,有无合理的小标题/分段
|
|
43
|
+
- **语言流畅性**:句子是否通顺,无冗余/歧义
|
|
44
|
+
- **篇幅适当性**:长度是否与内容密度匹配,无注水
|
|
45
|
+
|
|
46
|
+
**评分参考:**
|
|
47
|
+
- 9-10:标题抓人,结构一目了然,读来流畅,篇幅恰当
|
|
48
|
+
- 6-8:整体可读,有小问题(如段落过长、标题平淡)
|
|
49
|
+
- 3-5:结构混乱或语言拗口,影响阅读体验
|
|
50
|
+
- 1-2:难以阅读,结构缺失或大量错别字/语病
|
|
51
|
+
|
|
52
|
+
---
|
|
53
|
+
|
|
54
|
+
### 3. 叙事与钩子 Narrative Hook
|
|
55
|
+
|
|
56
|
+
文章是否有足够的叙事牵引力,让读者想读完并分享。
|
|
57
|
+
|
|
58
|
+
**Level 2 细则:**
|
|
59
|
+
- **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」
|
|
60
|
+
- **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
|
|
61
|
+
- **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
|
|
62
|
+
|
|
63
|
+
**评分参考:**
|
|
64
|
+
- 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
|
|
65
|
+
- 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)
|
|
66
|
+
- 3-5:以陈述为主,缺乏叙事钩子或工具化内容,金句稀少
|
|
67
|
+
- 1-2:纯说教型文章,无叙事、无工具、无金句
|
|
68
|
+
|
|
69
|
+
---
|
|
70
|
+
|
|
71
|
+
### 4. 可传播性 Shareability
|
|
72
|
+
|
|
73
|
+
文章是否具备被读者主动转发分享的动力。
|
|
74
|
+
|
|
75
|
+
**Level 2 细则:**
|
|
76
|
+
- **情感共鸣**:是否触动读者情绪(共鸣、惊喜、愤慨、感动等)
|
|
77
|
+
- **实用价值**:读者是否能从文章获得可直接使用的知识/工具
|
|
78
|
+
- **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
|
|
79
|
+
- **话题热度**:是否蹭了有传播势能的话题或事件
|
|
80
|
+
- **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
|
|
81
|
+
- **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
|
|
82
|
+
- **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
|
|
83
|
+
|
|
84
|
+
**评分参考:**
|
|
85
|
+
- 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
|
|
86
|
+
- 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)
|
|
87
|
+
- 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
|
|
88
|
+
- 1-2:缺乏传播动力,内容封闭,无 CTA,受众定位模糊
|
|
89
|
+
|
|
90
|
+
---
|
|
91
|
+
|
|
92
|
+
### 5. 时效性 Timeliness
|
|
93
|
+
|
|
94
|
+
文章内容是否与当前时间节点相关,发布时机是否恰当。
|
|
95
|
+
|
|
96
|
+
**Level 2 细则:**
|
|
97
|
+
- **话题新鲜度**:主题是否仍处于公众关注窗口内
|
|
98
|
+
- **发布时机**:相对于事件/节点,发布是领先、同步还是滞后
|
|
99
|
+
- **长尾价值**:若非即时性内容,是否具备长期参考价值
|
|
100
|
+
|
|
101
|
+
**评分参考:**
|
|
102
|
+
- 9-10:完美踩点,话题正热,发布时机极佳
|
|
103
|
+
- 6-8:时机尚可,话题热度略降或稍超前
|
|
104
|
+
- 3-5:话题冷却期发布,或时机明显滞后
|
|
105
|
+
- 1-2:内容严重过时,或与当下毫无关联
|
|
106
|
+
|
|
107
|
+
---
|
|
108
|
+
|
|
109
|
+
### 6. 事实准确性 Factual Accuracy
|
|
110
|
+
|
|
111
|
+
文章中的陈述、数据和逻辑是否经得起核查。
|
|
112
|
+
|
|
113
|
+
> ⚠️ 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。
|
|
114
|
+
|
|
115
|
+
**Level 2 细则:**
|
|
116
|
+
- **逻辑自洽**:论点与论据之间是否存在矛盾或跳跃
|
|
117
|
+
- **行业知识准确**:对行业概念、术语、常识的描述是否正确
|
|
118
|
+
- **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
|
|
119
|
+
- **来源可信度**:引用的信息来源是否权威、可溯源
|
|
120
|
+
|
|
121
|
+
**评分参考:**
|
|
122
|
+
- 9-10:逻辑严密,所有可查数据均已核实,行业知识准确
|
|
123
|
+
- 6-8:大体准确,有 1-2 处轻微错误或引用不严谨
|
|
124
|
+
- 3-5:多处存疑,数据或逻辑有明显问题
|
|
125
|
+
- 1-2:存在严重事实错误,或核心主张与可查来源矛盾
|
|
@@ -0,0 +1,90 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: article-to-image
|
|
3
|
+
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."
|
|
4
|
+
---
|
|
5
|
+
|
|
6
|
+
# Article-to-Image
|
|
7
|
+
|
|
8
|
+
Generate editorial cover images from article content. Three steps: analyze → brainstorm → generate.
|
|
9
|
+
|
|
10
|
+
## Format Selection
|
|
11
|
+
|
|
12
|
+
At the start of every invocation, read `references/formats.md` to get size, aspect ratio, title-safe zone, and prompt suffix for each format.
|
|
13
|
+
|
|
14
|
+
**Default format**: `wechat-banner` if not specified.
|
|
15
|
+
|
|
16
|
+
**Multi-format**: If the user requests multiple formats (e.g. "wechat + xiaohongshu"):
|
|
17
|
+
1. Designate the first-mentioned format as the **primary format**
|
|
18
|
+
2. Run the full brainstorm (Step 2) once for the primary format only — do NOT brainstorm per format
|
|
19
|
+
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)
|
|
20
|
+
|
|
21
|
+
## Universal Constraints (all formats)
|
|
22
|
+
|
|
23
|
+
- **Style**: Editorial illustration or cinematic photography — NOT cartoon, anime, or stock photo
|
|
24
|
+
- **Colors**: Max 3 dominant colors per image
|
|
25
|
+
- **No text in image**: Never embed title, captions, or article text
|
|
26
|
+
- **Depth**: Foreground/background separation — avoid flat compositions
|
|
27
|
+
- **Title-safe zone**: Always leave empty space per format spec for text overlay
|
|
28
|
+
|
|
29
|
+
## Workflow
|
|
30
|
+
|
|
31
|
+
### Step 1 — Extract core themes from article
|
|
32
|
+
|
|
33
|
+
Identify:
|
|
34
|
+
- **Central metaphor**: The one concept the article is really about (not the literal topic)
|
|
35
|
+
- **Emotional tone**: Intellectual curiosity / urgency / calm analysis / tension / etc.
|
|
36
|
+
- **Key visual anchors**: 2–3 concrete objects, scenes, or symbols implied by the article
|
|
37
|
+
- **Platform fit**: Does the tone match the requested format's style? (WeChat = serious; XHS = lifestyle)
|
|
38
|
+
|
|
39
|
+
### Step 2 — Brainstorm 4 visual directions
|
|
40
|
+
|
|
41
|
+
Each direction must differ in:
|
|
42
|
+
- Primary visual element (scene, object, symbol, environment)
|
|
43
|
+
- Compositional approach (tight portrait / wide environmental / abstract / conceptual)
|
|
44
|
+
- Emotional register
|
|
45
|
+
|
|
46
|
+
Present as a numbered list before generating. Format:
|
|
47
|
+
|
|
48
|
+
```
|
|
49
|
+
**Direction 1 — [Name]**
|
|
50
|
+
Visual: [specific scene description]
|
|
51
|
+
Mood: [emotional register]
|
|
52
|
+
Composition: [subject placement + title-safe side]
|
|
53
|
+
Style: [cinematic / editorial illustration / documentary]
|
|
54
|
+
```
|
|
55
|
+
|
|
56
|
+
Proceed to Step 3 immediately unless user asks to review first.
|
|
57
|
+
|
|
58
|
+
### Step 3 — Generate images
|
|
59
|
+
|
|
60
|
+
Build each prompt using this template:
|
|
61
|
+
|
|
62
|
+
```
|
|
63
|
+
[Specific scene: concrete subject, action, environment],
|
|
64
|
+
[lighting: named style e.g. "dramatic side lighting", "golden hour"],
|
|
65
|
+
[palette: name 2-3 colors],
|
|
66
|
+
[composition: e.g. "subject on right, left third empty for text overlay"],
|
|
67
|
+
[style: editorial photography / cinematic],
|
|
68
|
+
no text, no watermarks, no logos,
|
|
69
|
+
[format prompt suffix from formats.md]
|
|
70
|
+
```
|
|
71
|
+
|
|
72
|
+
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.
|
|
73
|
+
|
|
74
|
+
## Prompt Engineering Rules
|
|
75
|
+
|
|
76
|
+
**DO:**
|
|
77
|
+
- Use specific concrete scenes: "A taxi driver's hand refusing a banknote, warm amber interior"
|
|
78
|
+
- Name lighting explicitly: "golden hour side light", "cool blue overhead", "dramatic chiaroscuro"
|
|
79
|
+
- Reference photographic styles: "editorial photography", "documentary", "cinematic wide shot"
|
|
80
|
+
- Specify negative space: "empty left third for title overlay"
|
|
81
|
+
|
|
82
|
+
**DON'T:**
|
|
83
|
+
- Translate article topic literally into image (伊朗支付 ≠ 伊朗国旗 + 信用卡)
|
|
84
|
+
- Use vague adjectives alone: "beautiful", "professional", "modern"
|
|
85
|
+
- Include any text or UI elements
|
|
86
|
+
- Use styles that clash with the format's tone
|
|
87
|
+
|
|
88
|
+
## Format Reference
|
|
89
|
+
|
|
90
|
+
See `references/formats.md` for all supported formats with sizes, title-safe zones, and prompt suffixes.
|
|
@@ -0,0 +1,53 @@
|
|
|
1
|
+
# Output Format Specifications
|
|
2
|
+
|
|
3
|
+
## Supported Formats
|
|
4
|
+
|
|
5
|
+
### `wechat-banner` — 微信公众号首图
|
|
6
|
+
- **Size**: 900×383px (2.35:1)
|
|
7
|
+
- **Use case**: 文章顶部大图,分享卡片缩略图
|
|
8
|
+
- **Title-safe zone**: Left third empty OR top quarter empty
|
|
9
|
+
- **Style notes**: Editorial / cinematic, serious tone, industry-appropriate
|
|
10
|
+
- **Prompt suffix**: `wide format 2.35:1, professional WeChat article banner, editorial photography`
|
|
11
|
+
|
|
12
|
+
### `wechat-square` — 微信公众号方形封面
|
|
13
|
+
- **Size**: 900×900px (1:1)
|
|
14
|
+
- **Use case**: 公众号封面缩略图(列表页显示)
|
|
15
|
+
- **Title-safe zone**: Bottom quarter for title overlay
|
|
16
|
+
- **Style notes**: Strong single focal point, works at small size
|
|
17
|
+
- **Prompt suffix**: `square format 1:1, WeChat thumbnail cover, bold focal point`
|
|
18
|
+
|
|
19
|
+
### `xiaohongshu-cover` — 小红书封面
|
|
20
|
+
- **Size**: 1080×1350px (4:5)
|
|
21
|
+
- **Use case**: 小红书笔记封面图
|
|
22
|
+
- **Title-safe zone**: Top 20% and bottom 20% for text overlay
|
|
23
|
+
- **Style notes**: More lifestyle/aesthetic, warmer palette acceptable, slightly more vibrant than WeChat
|
|
24
|
+
- **Prompt suffix**: `portrait format 4:5, Xiaohongshu cover image, lifestyle editorial, vibrant but refined`
|
|
25
|
+
|
|
26
|
+
### `xiaohongshu-square` — 小红书方形
|
|
27
|
+
- **Size**: 1080×1080px (1:1)
|
|
28
|
+
- **Use case**: 小红书方形笔记封面
|
|
29
|
+
- **Title-safe zone**: Bottom third
|
|
30
|
+
- **Style notes**: Same as xiaohongshu-cover
|
|
31
|
+
- **Prompt suffix**: `square format 1:1, Xiaohongshu post cover, lifestyle editorial`
|
|
32
|
+
|
|
33
|
+
### `linkedin-banner` — LinkedIn 文章封面
|
|
34
|
+
- **Size**: 1200×627px (1.91:1)
|
|
35
|
+
- **Use case**: LinkedIn article header, link preview
|
|
36
|
+
- **Title-safe zone**: Center or left half
|
|
37
|
+
- **Style notes**: Corporate but not sterile, professional editorial
|
|
38
|
+
- **Prompt suffix**: `wide format 1.91:1, LinkedIn article banner, professional editorial photography`
|
|
39
|
+
|
|
40
|
+
### `general-wide` — 通用宽幅
|
|
41
|
+
- **Size**: 1920×1080px (16:9)
|
|
42
|
+
- **Use case**: Presentation slides, multi-platform reuse
|
|
43
|
+
- **Title-safe zone**: Left third or top quarter
|
|
44
|
+
- **Style notes**: Cinematic, high-impact
|
|
45
|
+
- **Prompt suffix**: `widescreen 16:9 format, cinematic, versatile editorial banner`
|
|
46
|
+
|
|
47
|
+
## Default
|
|
48
|
+
|
|
49
|
+
If no format is specified, default to `wechat-banner`.
|
|
50
|
+
|
|
51
|
+
## Adding New Formats
|
|
52
|
+
|
|
53
|
+
To add a new format, append a new section following the same structure: Size, Use case, Title-safe zone, Style notes, Prompt suffix.
|
|
@@ -0,0 +1,50 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: wechat-article-writer
|
|
3
|
+
description: "Write WeChat 公众号 articles in Chinese content styles — 踩坑型 (pitfall-story), 实操型 (how-to), 对比型 (comparison), 观察型 (industry insight). Use for drafting from scratch, rewriting a draft, adapting source material into a publishable article, or picking the right style; includes B2B/fintech guidance."
|
|
4
|
+
---
|
|
5
|
+
|
|
6
|
+
# WeChat Article Writer (公众号写作)
|
|
7
|
+
|
|
8
|
+
Write WeChat public account articles in proven Chinese content styles.
|
|
9
|
+
|
|
10
|
+
**Always read `references/writing-rules.md` before drafting** — it contains title formulas, language rules, brand voice guidance, and the pre-publish checklist.
|
|
11
|
+
|
|
12
|
+
## Step 1: Choose a Style
|
|
13
|
+
|
|
14
|
+
Ask the user which style to use, or recommend based on the topic. If unsure which style fits the topic, read `references/style-selector.md` for the selection framework and decision tree.
|
|
15
|
+
|
|
16
|
+
| Style | 适用场景 | Trigger keywords |
|
|
17
|
+
|-------|---------|-----------------|
|
|
18
|
+
| **踩坑型** | 经验教训、避坑指南、工具使用 | 踩坑、教训、搞懂了、终于明白 |
|
|
19
|
+
| **实操型** | 操作教程、配置指南、步骤分解 | 怎么做、教程、配置、攻略 |
|
|
20
|
+
| **对比型** | 概念澄清、产品对比、方案选型 | 区别、对比、哪个好、vs |
|
|
21
|
+
| **观察型** | 行业动态、趋势分析、深度观点 | 趋势、观察、为什么、背后 |
|
|
22
|
+
|
|
23
|
+
Read the style's reference file before drafting:
|
|
24
|
+
- 踩坑型 → `references/style-pitfall.md`
|
|
25
|
+
- 实操型 → `references/style-howto.md`
|
|
26
|
+
- 对比型 → `references/style-comparison.md`
|
|
27
|
+
- 观察型 → `references/style-insight.md`
|
|
28
|
+
|
|
29
|
+
## Step 2: Gather Inputs
|
|
30
|
+
|
|
31
|
+
Before writing, confirm:
|
|
32
|
+
1. **Topic** — What is the article about?
|
|
33
|
+
2. **Core insight** — What's the one thing the reader should take away?
|
|
34
|
+
3. **Source material** — Any research, links, or data to draw from?
|
|
35
|
+
4. **Target length** — Default: 800–1200 words (standard 公众号 length)
|
|
36
|
+
5. **Brand voice** — B2B professional, C-end consumer, or other? See `references/writing-rules.md` for guidance.
|
|
37
|
+
|
|
38
|
+
## Step 3: Draft the Article
|
|
39
|
+
|
|
40
|
+
Follow the structure from the chosen style's reference file. Apply these **non-negotiable requirements**:
|
|
41
|
+
|
|
42
|
+
- **Title** — Must use a formula from `references/writing-rules.md`; include a number or emotional hook
|
|
43
|
+
- **Opening hook (前200字)** — Must contain first-person narrative or a client case hook; readers must feel "this is about me" before reaching the body. No starting directly with definitions or background.
|
|
44
|
+
- **金句 (every 300–500 words)** — At least one standalone-quotable sentence per section, placed at paragraph start or end. Ending paragraph must close with a 金句. Types: analogy / counter-intuition / number-shock / self-deprecating (use sparingly for B2B)
|
|
45
|
+
- **CheckList or 行动步骤** — Every article must end with something actionable (checklist, numbered steps, or decision framework). Readers leave with a tool, not just information.
|
|
46
|
+
- **Closing CTA** — Invite next action appropriate to the brand (demo/consult for B2B; follow/share for C-end)
|
|
47
|
+
|
|
48
|
+
## Step 4: Self-check Before Delivering
|
|
49
|
+
|
|
50
|
+
Run the pre-publish checklist in `references/writing-rules.md` (发布前自查清单) before delivering.
|
|
@@ -0,0 +1,64 @@
|
|
|
1
|
+
# 风格:对比型(概念澄清 + 选型决策)
|
|
2
|
+
|
|
3
|
+
## 核心逻辑
|
|
4
|
+
|
|
5
|
+
「帮读者做决定,省去他们自己研究的时间。」
|
|
6
|
+
|
|
7
|
+
对比两个或多个选项,给出清晰的判断框架。读者的感受:「原来是这样,我知道该选哪个了。」
|
|
8
|
+
|
|
9
|
+
## 适用选题
|
|
10
|
+
|
|
11
|
+
- 概念澄清(支付失败 vs 拒付、预授权 vs 即时扣款)
|
|
12
|
+
- 方案选型(直连收单 vs 聚合支付、3DS vs 非3DS)
|
|
13
|
+
- 工具对比(不同支付网关的适用场景)
|
|
14
|
+
- 认知纠偏(「你以为的 X,实际上是 Y」)
|
|
15
|
+
|
|
16
|
+
## 文章结构
|
|
17
|
+
|
|
18
|
+
```
|
|
19
|
+
【标题】
|
|
20
|
+
[概念A] vs [概念B]:X% 的人分不清 / 你真的搞懂了吗?
|
|
21
|
+
示例:「支付失败 vs 交易拒付:90%的商家都搞混了,损失也因此扩大」
|
|
22
|
+
|
|
23
|
+
【开头(150字)】
|
|
24
|
+
抛出混淆现象:「很多商家碰到了 XXX,去做 YYY,结果发现没用」
|
|
25
|
+
说明为什么搞混了这两个概念会有实际损失
|
|
26
|
+
|
|
27
|
+
【概念澄清】
|
|
28
|
+
先用一句话分别定义两个概念:
|
|
29
|
+
「支付失败 = 交易没有发生;拒付 = 交易发生了但持卡人质疑」
|
|
30
|
+
|
|
31
|
+
【对比分析(核心)】
|
|
32
|
+
可以用维度对比,也可以用叙述式:
|
|
33
|
+
|
|
34
|
+
维度对比写法:
|
|
35
|
+
从 3–5 个角度分析两者的区别
|
|
36
|
+
每个维度:维度名 → A的情况 → B的情况 → 实际影响
|
|
37
|
+
|
|
38
|
+
叙述式写法:
|
|
39
|
+
讲一个真实场景,展示搞混后会发生什么
|
|
40
|
+
再讲正确处理流程对比
|
|
41
|
+
|
|
42
|
+
【决策框架】
|
|
43
|
+
「遇到 XXX 情况 → 判断是 A 还是 B 的方法」
|
|
44
|
+
给一个简单的判断树或口诀
|
|
45
|
+
|
|
46
|
+
【总结表格或要点】
|
|
47
|
+
用3–5条要点归纳,方便截图收藏
|
|
48
|
+
|
|
49
|
+
【结尾 CTA】
|
|
50
|
+
「如果你的系统还没区分这两类异常,[产品] 的风控模块可以帮你自动分类处理」
|
|
51
|
+
```
|
|
52
|
+
|
|
53
|
+
## 语气指南
|
|
54
|
+
|
|
55
|
+
- 开头可以直接点出「很多人搞错了」,不需要委婉
|
|
56
|
+
- 分析部分尽量客观,少用「最好」「一定」
|
|
57
|
+
- 结论部分给清晰建议,不要模棱两可
|
|
58
|
+
- B2B读者接受表格,但表格要简洁(不超过5列)
|
|
59
|
+
|
|
60
|
+
## 行业示例选题(B2B 支付/金融科技)
|
|
61
|
+
|
|
62
|
+
- 「支付失败 vs 交易拒付:商家必须搞清楚的两个概念」
|
|
63
|
+
- 「直连收单 vs 聚合支付:选错了就是在给自己挖坑」
|
|
64
|
+
- 「3DS 验证:做了真的更安全,还是只是增加了摩擦?」
|
|
@@ -0,0 +1,64 @@
|
|
|
1
|
+
# 风格:实操型(步骤教程 + 保姆级指南)
|
|
2
|
+
|
|
3
|
+
## 核心逻辑
|
|
4
|
+
|
|
5
|
+
「照着做就行。」
|
|
6
|
+
|
|
7
|
+
不需要读者理解原理,只需要跟着步骤走,最终得到结果。适合工具教程、操作指南、配置攻略。
|
|
8
|
+
|
|
9
|
+
## 适用选题
|
|
10
|
+
|
|
11
|
+
- 操作教程([产品] 风控规则配置、对账系统接入)
|
|
12
|
+
- 工具上手(API 集成步骤、后台操作流程)
|
|
13
|
+
- 问题排查(支付失败自查清单、拒付处理流程)
|
|
14
|
+
|
|
15
|
+
## 文章结构
|
|
16
|
+
|
|
17
|
+
```
|
|
18
|
+
【标题】
|
|
19
|
+
动作 + 保姆级/完整/手把手
|
|
20
|
+
示例:「手把手教你配置 A1 风控规则(附完整参数说明)」
|
|
21
|
+
|
|
22
|
+
【开头(150字)】
|
|
23
|
+
说清楚这篇文章解决什么问题:
|
|
24
|
+
「如果你在 XXX,读完这篇文章,你能做到 YYY」
|
|
25
|
+
可以加一句「很多人被卡在哪一步」建立共鸣
|
|
26
|
+
|
|
27
|
+
【前提条件(可选)】
|
|
28
|
+
列清楚需要准备什么,格式:
|
|
29
|
+
- 需要 XXX 权限
|
|
30
|
+
- 需要 XXX 文件/信息
|
|
31
|
+
- 预计耗时:XX 分钟
|
|
32
|
+
|
|
33
|
+
【步骤(主体)】
|
|
34
|
+
每步格式:
|
|
35
|
+
## 第一步:[动词 + 做什么]
|
|
36
|
+
[做什么,为什么这样做(简短)]
|
|
37
|
+
[具体操作,可以加截图/代码说明]
|
|
38
|
+
⚠️ 注意:[常见错误或易踩坑的点]
|
|
39
|
+
|
|
40
|
+
至少 3–5 步,每步独立、清晰
|
|
41
|
+
|
|
42
|
+
【常见问题 FAQ(可选)】
|
|
43
|
+
列出 3–5 个「做到这里卡住了怎么办」
|
|
44
|
+
|
|
45
|
+
【CheckList】
|
|
46
|
+
发布前/完成后的自查清单
|
|
47
|
+
|
|
48
|
+
【结尾】
|
|
49
|
+
「完成以上步骤,你应该能看到 XXX。如果还有问题,XXX」
|
|
50
|
+
CTA:「也可以联系我们,1对1帮你排查」
|
|
51
|
+
```
|
|
52
|
+
|
|
53
|
+
## 语气指南
|
|
54
|
+
|
|
55
|
+
- 每步动词开头:「打开 → 找到 → 填入 → 点击保存」
|
|
56
|
+
- 语气平实,不需要情绪
|
|
57
|
+
- 注意事项用 ⚠️ 标出,让读者不会漏掉
|
|
58
|
+
- FAQ 部分可以轻松一点:「你是不是被卡在这了?」
|
|
59
|
+
|
|
60
|
+
## 行业示例选题(B2B 支付/金融科技)
|
|
61
|
+
|
|
62
|
+
- 「[产品] 风控规则配置完整指南(从0到上线)」
|
|
63
|
+
- 「支付对账自动化:手把手接入 [产品] 对账 API」
|
|
64
|
+
- 「跨境收款失败自查清单:这7步排查覆盖90%的问题」
|
|
@@ -0,0 +1,63 @@
|
|
|
1
|
+
# 风格:观察型(行业洞察 + 深度观点)
|
|
2
|
+
|
|
3
|
+
## 核心逻辑
|
|
4
|
+
|
|
5
|
+
「我看到了一个别人没注意到的趋势/问题,告诉你它意味着什么。」
|
|
6
|
+
|
|
7
|
+
不卖课,不推产品,只分享观点和判断。适合建立行业权威感,培养长期读者忠诚度。
|
|
8
|
+
|
|
9
|
+
## 适用选题
|
|
10
|
+
|
|
11
|
+
- 行业趋势(跨境支付监管收紧、AI 风控兴起)
|
|
12
|
+
- 现象解读(为什么越来越多的商家在换支付网关)
|
|
13
|
+
- 深度分析(某政策/事件对支付行业的真实影响)
|
|
14
|
+
- 预判/展望(2025年支付行业会发生什么)
|
|
15
|
+
|
|
16
|
+
## 文章结构
|
|
17
|
+
|
|
18
|
+
```
|
|
19
|
+
【标题】
|
|
20
|
+
观察 + 为什么 / 背后 / 意味着什么
|
|
21
|
+
示例:「越来越多的跨境商家开始换支付网关——背后发生了什么?」
|
|
22
|
+
|
|
23
|
+
【开头(200字)】
|
|
24
|
+
抛出一个具体的现象或数据点,让读者觉得「哦,我也注意到了」
|
|
25
|
+
「最近我们接触到的商家里,有不少人在讨论 XXX」
|
|
26
|
+
「这个数字让我们觉得有点不寻常」
|
|
27
|
+
|
|
28
|
+
【现象拆解(2–3个视角)】
|
|
29
|
+
不是列事实,而是分析「为什么」
|
|
30
|
+
每个视角:
|
|
31
|
+
- 观察到的现象(1–2句)
|
|
32
|
+
- 背后的原因(分析,可以带数据)
|
|
33
|
+
- 这对谁有影响
|
|
34
|
+
|
|
35
|
+
避免:只堆行业新闻,不给观点
|
|
36
|
+
|
|
37
|
+
【我们的判断】
|
|
38
|
+
明确表达立场:「我们认为 XXX」
|
|
39
|
+
给出预测:「接下来6个月,我们预计会看到 XXX」
|
|
40
|
+
可以承认不确定性:「这个判断的前提是 XXX,如果 YYY 发生,结论可能不同」
|
|
41
|
+
|
|
42
|
+
【对读者的意义】
|
|
43
|
+
把行业趋势落到读者的实际操作上:
|
|
44
|
+
「如果你是 XXX,现在应该关注 YYY」
|
|
45
|
+
「具体来说,可以先做 ZZZ」
|
|
46
|
+
|
|
47
|
+
【结尾】
|
|
48
|
+
金句收尾,表达对行业的判断或态度
|
|
49
|
+
CTA:「我们很好奇你怎么看这个问题,欢迎留言」或「想了解 [产品] 怎么应对这个趋势?」
|
|
50
|
+
```
|
|
51
|
+
|
|
52
|
+
## 语气指南
|
|
53
|
+
|
|
54
|
+
- 开头可以用「我们」增加权威感
|
|
55
|
+
- 分析部分要有明确立场,不要「可能」「也许」全篇
|
|
56
|
+
- 但也要承认局限,不要过度自信
|
|
57
|
+
- B2B读者欣赏有深度的文章,不要怕文章稍长(1200–1500字可以)
|
|
58
|
+
|
|
59
|
+
## 行业示例选题(B2B 支付/金融科技)
|
|
60
|
+
|
|
61
|
+
- 「跨境支付监管在收紧:商家现在最需要做好的3件事」
|
|
62
|
+
- 「AI 风控为什么还没有大规模普及?我们拆了10个失败案例」
|
|
63
|
+
- 「支付行业的「中间层」正在消失——这对你意味着什么」
|