ReddRadar
Agent skill
screenshot-feature-extractor
Analyze product screenshots to extract feature lists and generate development task checklists. Use when: (1) Analyzing competitor product screenshots for feature extraction, (2) Generating PRD/task lists from UI designs, (3) Batch analyzing multiple app screens, (4) Conducting competitive analysis from visual references.
Install this agent skill to your Project
npx add-skill https://github.com/davila7/claude-code-templates/tree/main/cli-tool/components/skills/development/screenshot-feature-extractor
SKILL.md
Screenshot Analyzer (Multi-Agent)
Extract product features from UI screenshots using a coordinated multi-agent analysis pipeline.
Core principle: Describe WHAT to build (features/interactions), NOT HOW (no tech stack).
Multi-Agent Architecture
This skill orchestrates 5 specialized agents for comprehensive analysis:
┌─────────────────┐
│ Coordinator │
│ (this skill) │
└────────┬────────┘
│
┌───────────────────┼───────────────────┐
│ │ │
▼ ▼ ▼
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
│ UI Analyzer │ │ Interaction │ │ Business │
│ (parallel) │ │ Analyzer │ │ Analyzer │
│ │ │ (parallel) │ │ (parallel) │
└────────┬────────┘ └────────┬────────┘ └────────┬────────┘
│ │ │
└───────────────────┼───────────────────┘
▼
┌─────────────────┐
│ Synthesizer │
│ (sequential) │
└────────┬────────┘
│
▼
┌─────────────────┐
│ Reviewer │
│ (sequential) │
└─────────────────┘
Process
Phase 1: Screenshot Collection
Gather all screenshots to analyze:
- Read the screenshot file(s) provided by the user
- For each screenshot, note the file path and any context provided
- If multiple screenshots, determine if they are from the same product
Phase 2: Parallel Analysis
Launch THREE Task agents IN PARALLEL for each screenshot:
Agent 1: screenshot-ui-analyzer
Analyze this screenshot for UI components, layout structure, and design patterns.
Screenshot: [file path]
Return your analysis as JSON.
Agent 2: screenshot-interaction-analyzer
Analyze this screenshot for user interactions, navigation flows, and state transitions.
Screenshot: [file path]
Return your analysis as JSON.
Agent 3: screenshot-business-analyzer
Analyze this screenshot for business functions, data entities, and domain logic.
Screenshot: [file path]
Return your analysis as JSON.
IMPORTANT: Use the Task tool with THREE parallel calls in a single message to maximize efficiency.
Phase 3: Synthesis
After all parallel analyses complete, launch the synthesizer agent:
Agent 4: screenshot-synthesizer
Synthesize these analysis results into a unified development task list.
UI Analysis:
[paste UI analyzer result]
Interaction Analysis:
[paste Interaction analyzer result]
Business Analysis:
[paste Business analyzer result]
Product Name: [product name]
Output file: docs/plans/YYYY-MM-DD-<product>-features.md
Phase 4: Review
Launch the reviewer agent to validate the output:
Agent 5: screenshot-reviewer
Review this task list for completeness and quality.
Original screenshot(s): [file paths]
Task list: [synthesized output]
If issues found, provide corrections.
Phase 5: Output
- Write final task list to
docs/plans/YYYY-MM-DD-<product>-features.md - Use format from references/output-format.md
- Present summary to user
Key Guidelines
- Use
- [ ]checkbox format for all tasks - Break features into small, executable subtasks
- Focus on user interactions, not implementation details
- For multiple screenshots: deduplicate features across all screens
- For competitive analysis: highlight unique features and gaps
Benefits of Multi-Agent Approach
- Thoroughness - Three specialized perspectives catch more details
- Speed - Parallel analysis reduces total time
- Quality - Synthesis + Review ensures coherent, complete output
- Specialization - Each agent focuses on its domain expertise
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
davila7/claude-code-templates
verl-rl-training
Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.
davila7/claude-code-templates
openrlhf-training
High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.
davila7/claude-code-templates
gguf-quantization
GGUF format and llama.cpp quantization for efficient CPU/GPU inference. Use when deploying models on consumer hardware, Apple Silicon, or when needing flexible quantization from 2-8 bit without GPU requirements.
davila7/claude-code-templates
Claude Code Guide
Master guide for using Claude Code effectively. Includes configuration templates, prompting strategies "Thinking" keywords, debugging techniques, and best practices for interacting with the agent.
davila7/claude-code-templates
qdrant-vector-search
High-performance vector similarity search engine for RAG and semantic search. Use when building production RAG systems requiring fast nearest neighbor search, hybrid search with filtering, or scalable vector storage with Rust-powered performance.
davila7/claude-code-templates
behavioral-modes
AI operational modes (brainstorm, implement, debug, review, teach, ship, orchestrate). Use to adapt behavior based on task type.
Didn't find tool you were looking for?