Agent-Skills.md

Agent Skills: Photo Color

Adjust photo brightness, contrast, and saturation based on natural language prompts using GPT-5 text analysis. Use when the user asks to adjust brightness, change contrast, modify saturation, apply a specific style, or fix photo issues like washed out colors.

UncategorizedID: ppx123-web/claude-config/photo-color

Repository

ppx123-web/claude-config
ppx123-web

Install this agent skill to your local

pnpm dlx add-skill https://github.com/ppx123-web/claude-config/tree/HEAD/skills/photo-color

Skill Files

Browse the full folder contents for photo-color.

Download Skill

Loading file tree…

skills/photo-color/SKILL.md

Skill Metadata

Name
photo-color
Description
Adjust photo brightness, contrast, and saturation based on natural language prompts using GPT-5 text analysis. Use when the user asks to adjust brightness, change contrast, modify saturation, apply a specific style, or fix photo issues like washed out colors.

Photo Color

When to Use

Use this skill when the user asks to:

  • Adjust brightness (e.g., "make it brighter", "darken the photo")
  • Change contrast (e.g., "increase contrast", "make it more dramatic")
  • Modify saturation (e.g., "make colors more vibrant", "desaturate")
  • Apply a specific style (e.g., "cinematic look", "natural enhancement")
  • Fix photo issues (e.g., "fix washed out colors", "add warmth")

Instructions

This skill uses GPT-5 text API to interpret natural language and calculate color adjustments. Always:

  1. Validate input - Ensure photo exists and is JPG/PNG format
  2. Analyze photo - Calculate current brightness, contrast, saturation scores
  3. Call GPT-5 - Get enhancement suggestions based on user's prompt and photo analysis
  4. Select best match - Choose suggestion that best matches prompt keywords (vivid/natural/dramatic)
  5. Apply enhancements - Adjust colors and save to new file (never overwrite original)
  6. Report results - Show multipliers applied and GPT-5's reasoning

Usage

from photo_color import main

result = main(
    photo_path="photo.jpg",
    prompt="Make the colors more vibrant and warm"
)
# Returns: "photo-enhanced.jpg"

Command Line

python photo_color.py <photo_path> "<prompt>" [output_path]

Examples

Vivid colors:

Make the colors much more vibrant and vivid
Increase saturation by 50%
Make colors pop

Natural enhancement:

Apply a subtle natural enhancement
Keep it looking natural and soft

Cinematic/dramatic:

Create a dramatic cinematic look with high contrast
Make it bold and intense

Fix issues:

Fix the washed out appearance
Increase saturation to make colors more realistic
Add warmth to the photo

Specific adjustments:

Brighten and increase contrast
Make it darker and more moody

Output

  • Returns path to enhanced photo
  • Never overwrites original
  • Format: {original_name}-enhanced.{ext}
  • Multipliers in safe range (0.5-2.0)
  • Displays style name and GPT-5's reasoning

Enhancement Styles

GPT-5 suggests these style categories:

  • Vivid: High saturation (1.3-1.5), increased brightness and contrast
  • Natural: Subtle adjustments (1.05-1.2), minimal changes
  • Dramatic: High contrast (1.5-1.8), bold adjustments
  • Custom: Tailored to specific prompt requirements

Error Handling

  • Falls back to safe defaults (×1.2 saturation, ×1.1 brightness, ×1.1 contrast) if GPT-5 unavailable
  • Validates multipliers (0.5-2.0 range)
  • Clear error messages for missing files or invalid formats

Requirements

  • OPENROUTER_API_KEY environment variable must be set
  • Photo must be JPG or PNG format
  • Photo under 50MB recommended

Technical Implementation

See references/implementation.md for:

  • GPT-5 text API integration
  • Photo analysis algorithms
  • Color enhancement styles
  • Prompt matching strategy

Examples

See examples/common-use-cases.md for:

  • Vivid color enhancement
  • Natural subtle adjustment
  • Dramatic cinematic look
  • Fix washed out photos
  • Add warmth
  • Before/after comparisons