Notre avis
Ce compétence analyse l'entonnoir d'un quiz en ligne pour en extraire des informations concurrentielles, en capturant chaque écran et en générant un rapport détaillé.
Points forts
- Capture systématique de l'ensemble du parcours utilisateur dans un quiz
- Analyse approfondie des techniques de personnalisation et de vente
- Identification des tactiques psychologiques utilisées dans l'entonnoir
Limites
- Nécessite une inspection manuelle des captures d'écran pour certains détails
- Peut échouer sur des quiz utilisant une logique dynamique complexe ou des CAPTCHA
- Ne prend pas en charge les quiz nécessitant une authentification préalable
Utilisez cette compétence pour effectuer une analyse concurrentielle détaillée d'un quiz marketing et comprendre sa structure, sa personnalisation et son modèle de monétisation.
Évitez de l'utiliser si le quiz utilise un chiffrement côté client important ou si vous avez besoin d'une analyse en temps réel sans capture d'écran.
Analyse de sécurité
PrudenceThe skill instructs browser navigation, screenshot capture, and local file writing for competitive analysis. It does not exfiltrate data, execute remote scripts, or disable safety. However, the use of browser automation and Bash on user-provided URLs introduces risk if the user supplies malicious URLs. Overall, it is legitimate but powerful.
- •Uses browser automation tools to navigate arbitrary user-provided URLs, which could potentially access malicious sites.
- •Uses Bash to create directories and write files locally, but no exfiltration or network calls to external servers.
- •No explicit destructive commands, but the combination of powerful tools warrants caution.
Exemples
/analyze-quiz https://example.com/quiz/analyze-quiz https://example.com/quiz --name CompetitorNamename: analyze-quiz description: Analyze quiz funnels for competitive intelligence. Navigates through entire quiz, captures every screen, and generates comprehensive analysis report with structure, personalization, paywall analysis, and strengths/weaknesses.
Usage
/analyze-quiz <quiz-url>
/analyze-quiz <quiz-url> --name "CompanyName"
Instructions
You are analyzing a quiz funnel for competitive intelligence. Follow this process exactly.
Phase 1: Setup
- Extract company name from URL or
--nameparameter (default to domain name) - Create timestamp:
YYYY-MM-DD-HHmm - Create output directory using Bash:
mkdir -p ~/Downloads/quiz-analysis-[name]-[timestamp] - Initialize a working notes variable to track screen data
Phase 2: Quiz Navigation
Navigate to the quiz URL using browser_navigate.
For each screen, repeat until you reach the paywall/final offer:
-
Capture screenshot:
browser_take_screenshot with filename: ~/Downloads/quiz-analysis-[name]-[timestamp]/[##]-[screen-type].pngUse sequential numbering: 01, 02, 03... Screen types: landing, question, email-capture, loading, results, paywall, upsell
-
Capture accessibility snapshot:
browser_snapshotExtract: question text, answer options, button labels, progress indicators
-
Record screen data in your working notes:
- Screen number
- Question/content
- Answer options (if any)
- Question type (multiple choice, slider, text input, etc.)
- UI elements noted
-
Handle special inputs:
- Email field → type
test@example.com - Name field → type
Test User - Phone field → type
555-0100 - Age/number fields → use reasonable values (30, etc.)
- Email field → type
-
Navigate to next screen: Look for CTA buttons in this priority order:
- Explicit: "Next", "Continue", "Get Started", "Start Quiz", "Get Results", "See Results"
- Arrows: "→", "❯", right-arrow icons
- Submit buttons on forms
- Any prominent button that advances the flow
Use
browser_clickon the identified element. -
Wait for page transition: Use
browser_wait_forwith time: 1-2 seconds for animations -
Detect endpoint: Stop navigation when you see:
- Pricing/payment information
- Checkout form
- "Subscribe", "Buy", "Purchase" CTAs
- Plan comparison tables
- Credit card input fields
Phase 3: Analysis
After capturing all screens, analyze your collected data:
A. Quiz Structure
- Total number of screens
- Number of actual questions vs informational screens
- Question types used (list them)
- Progress indicator style (percentage, steps, bar, none)
- Estimated completion time
- Any branching/personalization indicators
B. Screen-by-Screen Summary Create a table: | # | Screenshot | Type | Content Summary | Options/Input | Notes | |---|------------|------|-----------------|---------------|-------|
C. Personalization Analysis
- What personal data is collected (list all fields)
- How is personalization promised to the user?
- "Calculating/analyzing" screen tactics (what do they show?)
- How does the results/offer page use collected data?
- Personalized copy patterns observed
D. Paywall/Offer Analysis
- Pricing structure (plans, prices, durations)
- Price anchoring tactics (crossed out prices, "was/now")
- Urgency/scarcity elements (timers, limited spots, etc.)
- Trial offers (free trial, money-back guarantee)
- Social proof on paywall (testimonials, user count, ratings)
- Trust signals (guarantees, security badges, payment icons)
- CTA button copy and design
- Multiple payment options shown?
- Upsells or downsells present?
E. Psychological Tactics Identified
- Commitment/consistency (small yeses leading to big yes)
- Social proof usage (where and how)
- Authority signals (experts, certifications, press logos)
- Scarcity/urgency (real or artificial)
- Loss aversion triggers
- Personalization/relevance hooks
- Sunk cost indicators (progress bars, time invested)
F. Key Strengths List 5-10 things they do well:
- Strong UX patterns
- Clever copy techniques
- Effective design choices
- Smart psychological triggers
- Worth adopting for your own quizzes
G. Key Weaknesses List 5-10 improvement opportunities:
- Friction points in the flow
- Missing trust elements
- Confusing or unclear screens
- Weak paywall elements
- Poor mobile considerations
- Missed personalization opportunities
Phase 4: Report Generation
Write the final markdown report to ~/Downloads/quiz-analysis-[name]-[timestamp].md:
# Quiz Funnel Analysis: [Company Name]
**URL:** [quiz-url]
**Analyzed:** [date]
**Total Screens:** [count]
## Executive Summary
[2-3 paragraph overview of the quiz, its approach, and key findings]
## Quiz Structure
[Section A findings]
## Screen-by-Screen Breakdown
[Section B table with screenshot references like ]
## Personalization Strategy
[Section C findings]
## Paywall & Offer Analysis
[Section D findings]
## Psychological Tactics
[Section E findings]
## Strengths
[Section F as bullet points]
## Weaknesses & Opportunities
[Section G as bullet points]
## Key Takeaways
[3-5 actionable insights for competitive advantage]
Output Confirmation
After generating the report, confirm to the user:
- Report location:
~/Downloads/quiz-analysis-[name]-[timestamp].md - Screenshots folder:
~/Downloads/quiz-analysis-[name]-[timestamp]/ - Number of screens captured
- Brief summary of key findings
Error Handling
- If quiz requires account/login: Stop and report, note what screens were captured
- If quiz has CAPTCHA: Stop and report
- If navigation gets stuck: Try alternative button selectors, report if still stuck
- If paywall requires payment to continue: Stop at paywall (expected behavior)
Tips
- Some quizzes have loading/animation screens - wait for them to complete
- Watch for A/B test indicators in URLs (variant=, ab=, etc.) and note them
- If options seem to branch the quiz, note which path was taken
- Capture any exit-intent popups or special offers that appear
Ingénierie de Prompts
Data & IA
Bonnes pratiques et templates de prompt engineering pour maximiser les résultats IA.
Visualisation de Données
Data & IA
Génère des visualisations de données et graphiques adaptés à vos données.
Architecture RAG
Data & IA
Guide de configuration d'architectures RAG (Retrieval-Augmented Generation).