GIF Analyzer

Analyze animated GIFs by extracting frames and interpreting them as sequential video.

Quick Start

When the user uses /gif:

Pattern 1: Basic analysis

/gif ./animation.gif

Pattern 2: With specific request

/gif ./animation.gif describe what happens step by step
/gif ./demo.gif what is the character doing?
/gif ./screen.gif summarize this screen recording

Workflow

Step 1: Parse User Input

Extract from the user's message:

1. GIF path: The file path immediately after /gif
2. User request (optional): Any text after the GIF path

Example parsing:

• Input: /gif ./demo.gif explain the animation
• GIF path: ./demo.gif
• User request: explain the animation

Step 2: Extract Frames

python3 ~/.claude/skills/gif-analyzer/extract_gif_frames.py <gif_path> --output-dir /tmp/gif_frames_analysis

Options for long GIFs:

python3 ~/.claude/skills/gif-analyzer/extract_gif_frames.py <gif_path> --max-frames 30 --skip 2

Step 3: Read Metadata

Check gif_metadata.json for:

• Total frames and duration
• Individual frame timestamps
• Resolution and loop information

Step 4: View and Analyze Frames

View frames in order: frame_001.png, frame_002.png, etc.

CRITICAL: Treat frames as a continuous video sequence:

1. Frame 001 = START of the animation
2. Frame numbers increase in TIME ORDER
3. Consecutive frames show MOTION/CHANGE over time

Step 5: Respond to User

If user provided a specific request, focus on answering that. Otherwise, provide a general analysis of the GIF content.

Script Options

Option	Description	Default
--output-dir, -o	Output directory	./gif_frames_<timestamp>
--max-frames, -m	Max frames to extract	50
--skip, -s	Extract every Nth frame	1 (all frames)

Frame Analysis Guidelines

1. Temporal Awareness: Frame 001 is the beginning
2. Motion Detection: Compare adjacent frames to identify movement
3. Key Frames: Identify significant moments (start, middle, end)
4. Loop Points: Note if the animation appears to loop
5. Duration Context: Use timestamp info to understand pacing

Output Format

📹 **GIF Analysis: [filename]**

**Overview:**
[1-2 sentence summary]

**Timeline:**
- [0.0s - 0.5s] Frame 1-5: [Description]
- [0.5s - 1.0s] Frame 6-10: [Description]
...

**[Answer to user's specific question if provided]**

Troubleshooting

• Pillow not found: pip install Pillow
• Too many frames: Use --skip 2 or higher
• Large output: Use --max-frames 20