xhs-image-gen

Name: xhs-image-gen
Author: PHY041

18stars2forks

How to install xhs-image-gen?

phy041/claude-agent-skills/xhs-image-gen

$npx -y skills add PHY041/claude-agent-skills --skill xhs-image-gen

Installs into the current project.

Summary

Generate Xiaohongshu (小红书) carousel images (3:4, 1080x1440) from structured content. Creates professional slide decks with Chinese typography and proper XHS formatting. Triggers on "做小红书图", "xhs images", "generate xhs slides", "小红书配图", "carousel images", or any XHS image generati

Files · 1

View on GitHub

SKILL.md

1	# XHS Image Generator Skill
2
3	Generate professional Xiaohongshu carousel images from structured content JSON. Outputs 3:4 vertical PNGs (1080x1440) with Chinese typography, ready for upload.
4
5	---
6
7	## Setup
8
9	Clone the skill:
10	```bash
11	git clone https://github.com/PHY041/claude-agent-skills
12	cd claude-agent-skills/skills/xhs-image-gen
13	python3 -m venv venv
14	venv/bin/pip install -r requirements.txt
15	```
16
17	Optional — AI cover generation (uses fal.ai):
18	```bash
19	export FAL_API_KEY="your_fal_api_key"
20	```
21
22	---
23
24	## Quick Start
25
26	```bash
27	# Template-only (fast, free)
28	venv/bin/python3 generate.py content.json --output /tmp/xhs-output/
29
30	# AI cover + template content (recommended)
31	FAL_API_KEY="..." venv/bin/python3 generate.py content.json --ai-cover --output /tmp/xhs-output/
32	```
33
34	---
35
36	## Workflow
37
38	### Step 1: Understand the Topic
39
40	When user asks to create XHS content for topic X:
41	1. Understand the topic and target audience
42	2. Plan 4-6 slides (cover + 2-4 content + summary)
43	3. Choose appropriate template
44
45	### Step 2: Decide on Cover Style
46
47	\| Mode \| Flag \| When to Use \|
48	\|------\|------\|-------------\|
49	\| Gradient cover (default) \| _(none)_ \| Clean, fast, no API cost \|
50	\| AI cover \| `--ai-cover` \| Eye-catching background via fal.ai \|
51
52	AI cover uses flux/schnell to generate an atmospheric background, applies a dark gradient overlay, and renders white text on top. Content slides always use templates (no AI). Cost: ~$0.003 per cover.
53
54	### Step 3: Generate Content JSON
55
56	```json
57	{
58	"template": "tutorial",
59	"ai_cover": true,
60	"slides": [
61	{
62	"type": "cover",
63	"title": "大标题（简短有力）",
64	"subtitle": "副标题（补充说明）",
65	"tag": "分类标签",
66	"author": "@your_handle",
67	"ai_prompt": "optional: custom fal.ai prompt for background"
68	},
69	{
70	"type": "content",
71	"title": "第一个要点",
72	"number": 1,
73	"body": "详细说明\n用换行分段\n每段1-2句话",
74	"highlight": "可选：重点提示框内容"
75	},
76	{
77	"type": "list",
78	"title": "清单/工具列表",
79	"items": [
80	"第一项 — 简短说明",
81	"第二项 — 简短说明",
82	"第三项 — 简短说明"
83	]
84	},
85	{
86	"type": "summary",
87	"title": "总结",
88	"points": [
89	"核心收获一",
90	"核心收获二",
91	"核心收获三"
92	],
93	"cta": "关注我获取更多干货"
94	}
95	]
96	}
97	```
98
99	### Step 4: Generate Images
100
101	```bash
102	cat > /tmp/xhs-content.json << 'EOF'
103	{... your JSON here ...}
104	EOF
105
106	venv/bin/python3 generate.py /tmp/xhs-content.json --output /tmp/xhs-output/
107	```
108
109	### Step 5: Present to User
110
111	Show the generated images for review. If approved, they're ready for manual upload to Xiaohongshu.
112
113	---
114
115	## Available Templates
116
117	\| Template \| Style \| Best For \|
118	\|----------\|-------\|----------\|
119	\| `tutorial` \| White bg, blue accent \| How-to, tutorials, step-by-step \|
120	\| `dark` \| Dark bg, soft blue accent \| Tech, dev tools, coding \|
121	\| `warm` \| Cream bg, amber accent \| Lifestyle, personal stories \|
122	\| `nature` \| Light green bg, green accent \| Wellness, growth, nature \|
123	\| `coral` \| Light pink bg, red accent \| Alerts, hot takes, comparisons \|
124
125	Select with `--template dark` or set `"template": "dark"` in JSON.
126
127	### Custom Accent Color
128
129	Override with `"accent_color": "#FF6B6B"` in the JSON root.
130
131	---
132
133	## Content Rules (CRITICAL)
134
135	### Slide Count
136	- Minimum 3 slides (cover + 1 content + summary)
137	- Optimal 4-6 slides (XHS algorithm favors multi-image carousels)
138	- Maximum 9 slides (beyond this, completion rate drops)
139
140	### Text Length Per Slide
141	- Title: 6-12 characters (Chinese) — must be scannable
142	- Body: 4-6 lines max per slide — don't cram
143	- List items: 8 max per slide
144	- CTA: Under 12 characters
145
146	### Language & Tone
147	- Write in Chinese (simplified)
148	- 朋友分享 tone — like telling a friend, not marketing
149	- Short sentences, line breaks between ideas
150	- Concrete numbers > vague claims ("省90%时间" > "节省大量时间")
151
152	### What NOT to Do
153	- No English-only slides (mixed OK, pure English bad for XHS)
154	- No walls of text (if body > 6 lines, split into 2 slides)
155	- No generic AI-sounding language
156	- No hype words: 颠覆、革命性、10倍、全网最强
157
158	---
159
160	## CLI Options
161
162	```
163	generate.py <content.json> [--template NAME] [--output DIR] [--ai-cover]
164	generate.py --list-templates
165	```
166
167	\| Flag \| Effect \|
168	\|------\|--------\|
169	\| `--ai-cover` \| Generate fal.ai background for cover slide \|
170	\| `--template` \| Template preset (tutorial, dark, warm, nature, coral) \|
171	\| `--output` \| Output directory (default: ./outp

People who liked this also viewed

6 picks

Type

xhs-image-gen

18stars2forks

View on GitHub

How to install xhs-image-gen?

phy041/claude-agent-skills/xhs-image-gen

$npx -y skills add PHY041/claude-agent-skills --skill xhs-image-gen

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# XHS Image Generator Skill
2
3	Generate professional Xiaohongshu carousel images from structured content JSON. Outputs 3:4 vertical PNGs (1080x1440) with Chinese typography, ready for upload.
4
5	---
6
7	## Setup
8
9	Clone the skill:
10	```bash
11	git clone https://github.com/PHY041/claude-agent-skills
12	cd claude-agent-skills/skills/xhs-image-gen
13	python3 -m venv venv
14	venv/bin/pip install -r requirements.txt
15	```
16
17	Optional — AI cover generation (uses fal.ai):
18	```bash
19	export FAL_API_KEY="your_fal_api_key"
20	```
21
22	---
23
24	## Quick Start
25
26	```bash
27	# Template-only (fast, free)
28	venv/bin/python3 generate.py content.json --output /tmp/xhs-output/
29
30	# AI cover + template content (recommended)
31	FAL_API_KEY="..." venv/bin/python3 generate.py content.json --ai-cover --output /tmp/xhs-output/
32	```
33
34	---
35
36	## Workflow
37
38	### Step 1: Understand the Topic
39
40	When user asks to create XHS content for topic X:
41	1. Understand the topic and target audience
42	2. Plan 4-6 slides (cover + 2-4 content + summary)
43	3. Choose appropriate template
44
45	### Step 2: Decide on Cover Style
46
47	\| Mode \| Flag \| When to Use \|
48	\|------\|------\|-------------\|
49	\| Gradient cover (default) \| _(none)_ \| Clean, fast, no API cost \|
50	\| AI cover \| `--ai-cover` \| Eye-catching background via fal.ai \|
51
52	AI cover uses flux/schnell to generate an atmospheric background, applies a dark gradient overlay, and renders white text on top. Content slides always use templates (no AI). Cost: ~$0.003 per cover.
53
54	### Step 3: Generate Content JSON
55
56	```json
57	{
58	"template": "tutorial",
59	"ai_cover": true,
60	"slides": [
61	{
62	"type": "cover",
63	"title": "大标题（简短有力）",
64	"subtitle": "副标题（补充说明）",
65	"tag": "分类标签",
66	"author": "@your_handle",
67	"ai_prompt": "optional: custom fal.ai prompt for background"
68	},
69	{
70	"type": "content",
71	"title": "第一个要点",
72	"number": 1,
73	"body": "详细说明\n用换行分段\n每段1-2句话",
74	"highlight": "可选：重点提示框内容"
75	},
76	{
77	"type": "list",
78	"title": "清单/工具列表",
79	"items": [
80	"第一项 — 简短说明",
81	"第二项 — 简短说明",
82	"第三项 — 简短说明"
83	]
84	},
85	{
86	"type": "summary",
87	"title": "总结",
88	"points": [
89	"核心收获一",
90	"核心收获二",
91	"核心收获三"
92	],
93	"cta": "关注我获取更多干货"
94	}
95	]
96	}
97	```
98
99	### Step 4: Generate Images
100
101	```bash
102	cat > /tmp/xhs-content.json << 'EOF'
103	{... your JSON here ...}
104	EOF
105
106	venv/bin/python3 generate.py /tmp/xhs-content.json --output /tmp/xhs-output/
107	```
108
109	### Step 5: Present to User
110
111	Show the generated images for review. If approved, they're ready for manual upload to Xiaohongshu.
112
113	---
114
115	## Available Templates
116
117	\| Template \| Style \| Best For \|
118	\|----------\|-------\|----------\|
119	\| `tutorial` \| White bg, blue accent \| How-to, tutorials, step-by-step \|
120	\| `dark` \| Dark bg, soft blue accent \| Tech, dev tools, coding \|
121	\| `warm` \| Cream bg, amber accent \| Lifestyle, personal stories \|
122	\| `nature` \| Light green bg, green accent \| Wellness, growth, nature \|
123	\| `coral` \| Light pink bg, red accent \| Alerts, hot takes, comparisons \|
124
125	Select with `--template dark` or set `"template": "dark"` in JSON.
126
127	### Custom Accent Color
128
129	Override with `"accent_color": "#FF6B6B"` in the JSON root.
130
131	---
132
133	## Content Rules (CRITICAL)
134
135	### Slide Count
136	- Minimum 3 slides (cover + 1 content + summary)
137	- Optimal 4-6 slides (XHS algorithm favors multi-image carousels)
138	- Maximum 9 slides (beyond this, completion rate drops)
139
140	### Text Length Per Slide
141	- Title: 6-12 characters (Chinese) — must be scannable
142	- Body: 4-6 lines max per slide — don't cram
143	- List items: 8 max per slide
144	- CTA: Under 12 characters
145
146	### Language & Tone
147	- Write in Chinese (simplified)
148	- 朋友分享 tone — like telling a friend, not marketing
149	- Short sentences, line breaks between ideas
150	- Concrete numbers > vague claims ("省90%时间" > "节省大量时间")
151
152	### What NOT to Do
153	- No English-only slides (mixed OK, pure English bad for XHS)
154	- No walls of text (if body > 6 lines, split into 2 slides)
155	- No generic AI-sounding language
156	- No hype words: 颠覆、革命性、10倍、全网最强
157
158	---
159
160	## CLI Options
161
162	```
163	generate.py <content.json> [--template NAME] [--output DIR] [--ai-cover]
164	generate.py --list-templates
165	```
166
167	\| Flag \| Effect \|
168	\|------\|--------\|
169	\| `--ai-cover` \| Generate fal.ai background for cover slide \|
170	\| `--template` \| Template preset (tutorial, dark, warm, nature, coral) \|
171	\| `--output` \| Output directory (default: ./outp

People who liked this also viewed

6 picks

Type