Skill 详情

brand-research-agent

Researches an existing brand for content production rather than creating one.

匹配类型可能匹配已针对 品牌打造 审核
来源michaelboeding/skills外部来源
报告安装量29仅表示受欢迎程度

使用前先检查

自动化审核只检查相关性,不代表安全审查或推荐。使用前请阅读来源中的说明。

已保存的来源预览

SKILL.md

这段内容是审核时保存的快照。外部来源才是完整且最新的版本。

---
name: brand-research-agent
description: >
  Use this skill to analyze a brand from their website before creating content.
  Triggers: "analyze brand", "research brand", "brand guidelines", "brand profile",
  "understand brand", "brand colors", "brand voice", "before creating content for"
  Extracts: colors, typography, voice/tone, products, audience, competitive positioning.
  Outputs a reusable brand_profile.json that producer skills can use.
---

# Brand Research Skill

Analyze a brand's website to extract brand elements for content creation.

**This skill uses 5 specialized agents** that analyze different brand dimensions in parallel, then synthesizes into a unified brand profile.

## What It Extracts

| Dimension | Elements | Used For |
|-----------|----------|----------|
| **Visual** | Colors, typography, logo, imagery style | Image prompts, video style |
| **Voice** | Tone, messaging, taglines, copy style | Voiceover, scripts |
| **Product** | Offerings, features, USPs, pricing | Content focus |
| **Audience** | Demographics, psychographics, pain points | Tone targeting |
| **Positioning** | Market position, competitors, differentiation | Messaging strategy |

## Prerequisites

- Browser access (uses web scraping)
- No API keys required

## Workflow

### Step 1: Get the Website URL (REQUIRED)

**FIRST, always ask for the website URL.** Do not proceed without it.

**Say this to the user:**

> "I'll analyze that brand for you! 
>
> **What's the website URL?** (e.g., https://nike.com)
>
> Optional: Any specific pages to focus on? Any known competitors?"

**Wait for the user to provide the URL before proceeding.**

If the user just says "analyze Nike's brand", ask:
> "I'll analyze Nike for you. Just to confirm - should I use https://nike.com?"

---

### Step 2: Navigate and Capture Website Data

Use browser tools to navigate the website and capture:

```
1. Homepage - Overall impression, hero messaging, colors
2. About page - Company story, values, team
3. Products/Services page - Offerings, features, pricing
4. Blog/Resources - Voice, content style
5. Footer - Social links, taglines
```

**Capture for each page:**
- Screenshot for visual analysis
- HTML/CSS for colors and fonts
- Text content for voice analysis
- Images for style analysis

---

### Step 3: Run Specialized Agents in Parallel

Deploy 5 agents, each analyzing a different dimension:

#### Agent 1: Visual Analyst
Focus: Colors, typography, logo, imagery style
```
Analyze:
- Primary, secondary, accent colors from CSS/design
- Font families for headings and body
- Logo usage and placement
- Photography/illustration style
- Overall visual mood (minimal, bold, playful, etc.)
```

#### Agent 2: Voice Analyst
Focus: Tone, messaging, taglines, copy style
```
Analyze:
- Headlines and how they're written
- Body copy style (formal vs casual)
- CTAs and their tone
- Taglines and slogans
- Emotional appeals used
```

#### Agent 3: Product Analyst
Focus: Offerings, features, USPs, pricing
```
Analyze:
- What products/services they offer
- Key features highlighted
- Unique selling propositions
- Pricing model and tiers
- Value propositions
```

#### Agent 4: Audience Analyst
Focus: Demographics, psychographics, pain points
```
Analyze:
- Who the messaging targets
- Pain points addressed
- Aspirations appealed to
- Language level and jargon
- Testimonials and case studies
```

#### Agent 5: Competitive Analyst
Focus: Market position, competitors, differentiation
```
Analyze:
- Market/industry category
- Implied competitors
- How they differentiate
- Unique market position
- Competitive advantages claimed
```

---

### Step 4: Synthesize into Brand Profile

Combine all agent outputs into a unified `brand_profile.json`:

```json
{
  "brand": {
    "name": "Company Name",
    "website": "https://example.com",
    "tagline": "Their main tagline",
    "analyzed_date": "2026-01-04"
  },
  "visual": {
    "colors": {
      "primary": "#1E40AF",
      "secondary": "#F59E0B",
     
在 GitHub 阅读完整来源 (打开外部页面)
相关上下文

相关工作