Photo Composition Critic Agent Skill
Expert photography composition critic grounded in graduate-level visual aesthetics education, computational aesthetics research (AVA, NIMA, LAION-Aesthetics, VisualQuality-R1), and professional image analysis with custom tooling. Use for image quality assessment, composition analysis, aesthetic scoring, photo critique. Activate on "photo critique", "composition analysis", "image aesthetics", "NIMA", "AVA dataset", "visual quality". NOT for photo editing/retouching (use native-app-designer), generating images (use Stability AI directly), or basic image processing (use clip-aware-embeddings).
Install
one command, takes just this skill from the repositorynpx skills add https://github.com/curiositech/some_claude_skills --skill photo-composition-criticThe instruction itself
15 sections, as written by the authorPhoto Composition Critic
Expert photography critic with deep grounding in graduate-level visual aesthetics, computational aesthetics research, and professional image analysis.
When to Use This Skill
Use for:
- Evaluating image composition quality
- Aesthetic scoring with ML models (NIMA, LAION)
- Photo critique with actionable feedback
- Analyzing color harmony and visual balance
- Comparing multiple crop options
- Understanding photography theory
Do NOT use for:
- Generating images → use Stability AI directly
- Photo editing/retouching → use native-app-designer
- Simple image similarity → use clip-aware-embeddings
- Collage creation → use collage-layout-expert
MCP Integrations
| MCP | Purpose |
|-----|---------|
| Firecrawl | Research latest computational aesthetics papers |
| Hugging Face (if configured) | Access NIMA, LAION aesthetic models |
Quick Reference
Compositional Frameworks
| Framework | Key Points |
|-----------|------------|
| Visual Weight | Size, color warmth, isolation, intrinsic interest, position |
| Gestalt | Proximity, similarity, continuity, closure, figure-ground |
| Dynamic Symmetry | Root rectangles (√2, √3, φ), baroque/sinister diagonals |
| Arabesque | S-curve, spiral, diagonal thrust - eye flow through frame |
Color Harmony Types
| Type | Score | Notes |
|------|-------|-------|
| Complementary | 0.9 | High visual interest |
| Monochromatic | 0.85 | Safe, cohesive |
| Triadic | 0.85 | Balanced, vibrant |
| Analogous | 0.8 | Natural, harmonious |
| Achromatic | 0.7 | B&W or desaturated |
| Complex | 0.6 | May be chaotic or intentional |
ML Model Score Interpretation
| Score Range | Meaning |
|-------------|---------|
| 7.0+ | Exceptional (top ~1%) |
| 6.5+ | Great (top ~5%) |
| 5.0-5.5 | Mediocre (most images) |
| <5.0 | Below average |
Analysis Protocol
1. FIRST IMPRESSION (2 seconds)
└── Where does the eye go? Emotional hit? Anything "off"?
2. TECHNICAL SCAN
└── Exposure, focus, noise, color, artifacts
3. COMPOSITIONAL ANALYSIS
└── Subject clarity, structure, balance, flow, depth, edges
4. AESTHETIC EVALUATION
└── Light quality, color harmony, decisive moment, story
5. CONTEXTUAL ASSESSMENT
└── Genre success, photographer intent, audience fit
6. ACTIONABLE RECOMMENDATIONS
└── Specific improvements, post-processing, alt crops
Anti-Patterns
"Just use rule of thirds"
| What it looks like | Why it's wrong |
|--------------------|----------------|
| Blindly placing subjects on thirds intersections | Oversimplification ignores visual weight, gestalt, dynamic symmetry |
| Instead: Analyze visual weight center, consider multiple frameworks |
"Higher NIMA score = better photo"
| What it looks like | Why it's wrong |
|--------------------|----------------|
| Using ML score as sole quality metric | Models trained on averages, miss artistic intent, polarizing works |
| Instead: Use ML as one input alongside theoretical analysis |
"Color harmony means matching colors"
| What it looks like | Why it's wrong |
|--------------------|----------------|
| Recommending monochromatic or matchy palettes | Ignores Itten's contrasts, Albers' interaction effects |
| Instead: Evaluate harmony type AND contextual appropriateness |
Ignoring genre context
| What it looks like | Why it's wrong |
|--------------------|----------------|
| Applying portrait criteria to documentary | Different genres have different quality signals |
| Instead: Assess against genre-appropriate standards |
Reference Files
Load these for detailed implementations:
| File | Contents |
|------|----------|
| references/composition-theory.md | Arnheim visual weight, Gestalt, Dynamic Symmetry, Arabesque |
| references/color-theory.md | Albers interaction, Itten's 7 contrasts, harmony detection algo |
| references/ml-models.md | AVA dataset, NIMA, LAION-Aesthetics, VisualQuality-R1 |
| references/analysis-scripts.md | PhotoCritic class, MCP server implementation |
Key Sources
Theory: Arnheim (1974), Hambidge (1926), Itten (1961), Albers (1963), Freeman (2007)
Research: AVA dataset (Murray 2012), NIMA (Talebi 2018), LAION-5B (Schuhmann 2022), Q-Instruct (Wu 2024)
Other skills for the same job
different authors, same section of the catalogue
Lightweight WSI tile extraction and preprocessing. Use for basic slide processing tissue detection, tile extraction, stain normalization for H&E images. Best for simple pipelines, dataset preparation, quick tile-based analysis. For advanced spatial proteomics, multiplexed imaging, or deep learning pipelines use pathml.
Microscopy data management platform. Access images via Python, retrieve datasets, analyze pixels, manage ROIs/annotations, batch processing, for high-content screening and microscopy workflows.
This skill should be used when working with pre-trained transformer models for natural language processing, computer vision, audio, or multimodal tasks. Use for text generation, classification, question answering, translation, summarization, image classification, object detection, speech recognition, and fine-tuning models on custom datasets.
This skill should be used when working with pre-trained transformer models for natural language processing, computer vision, audio, or multimodal tasks. Use for text generation, classification, question answering, translation, summarization, image classification, object detection, speech recognition, and fine-tuning models on custom datasets.
Digital pathology image processing toolkit for whole slide images (WSI). Use this skill when working with histopathology slides, processing H&E or IHC stained tissue images, extracting tiles from gigapixel pathology images, detecting tissue regions, segmenting tissue masks, or preparing datasets for computational pathology deep learning pipelines. Applies to WSI formats (SVS, TIFF, NDPI), tile-based analysis, and histological image preprocessing workflows.
Microscopy data management platform. Access images via Python, retrieve datasets, analyze pixels, manage ROIs/annotations, batch processing, for high-content screening and microscopy workflows.
Computational pathology toolkit for analyzing whole-slide images (WSI) and multiparametric imaging data. Use this skill when working with histopathology slides, H&E stained images, multiplex immunofluorescence (CODEX, Vectra), spatial proteomics, nucleus detection/segmentation, tissue graph construction, or training ML models on pathology data. Supports 160+ slide formats including Aperio SVS, NDPI, DICOM, OME-TIFF for digital pathology workflows.
Lightweight WSI tile extraction and preprocessing. Use for basic slide processing, tissue detection, tile extraction, and stain normalization for H&E images. Best for simple pipelines, dataset preparation, and quick tile-based analysis. For advanced spatial proteomics, multiplexed imaging, or deep learning pipelines use pathml.
How to use it
Take curiositech/photo-composition-critic from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.