Photos

What This Skill Does

The Photos skill acts as an intelligent, privacy-conscious curator for your local image repositories. It transforms static photo collections into a searchable, organized database using non-destructive metadata handling. By generating perceptual hashes and leveraging local vision analysis, the skill enables users to find specific memories without relying on proprietary cloud services. It features an advanced indexing system that stores metadata in separate JSON sidecars, ensuring your original files remain untouched and compatible with any other software. Whether you are managing thousands of disorganized images or looking to reclaim storage space, this skill provides safe automation for duplicate detection, chronological renaming, and semantic searching.

Installation

To integrate this skill into your environment, run the following command in your terminal: clawhub install openclaw/skills/skills/ivangdavila/photos

Use Cases

• Bulk Cleanup: Automatically identify near-duplicates and resized copies to save disk space without accidentally deleting unique shots.
• Semantic Search: Locate photos by describing content, such as "show me photos of a cat on a wooden table," rather than relying on folder structures.
• Archive Management: Rename disorganized piles of photos into a structured YYYY/MM/DD hierarchy with sortable filenames.
• Privacy-First Sharing: Prepare images for social media by automatically stripping GPS data while keeping your source files intact.

Example Prompts

1. "Find all photos from my 2023 summer trip that include the beach and create a folder named 'Summer 2023' with copies."
2. "Scan my photography folder for duplicates and show me a list of candidates to move to the trash; I only want to keep the highest resolution versions."
3. "Rename all images in the 'Downloads' folder to use the format YYYYMMDD_HHMMSS and group them by month."

Tips & Limitations

• Non-Destructive Defaults: The skill follows a 'safety first' philosophy. It never deletes files directly; instead, it moves them to a .photo-trash/ folder. Always review the photos-pending.json manifest before executing bulk operations.
• Indexing Performance: Indexing happens incrementally. The first pass on a large library may take time depending on your hardware, but subsequent runs are significantly faster as the skill caches results in sidecars.
• Vision Analysis: To optimize performance and privacy, vision analysis is performed on-demand. If the AI doesn't recognize an object immediately, the metadata is updated once a specific search confirms the need for analysis.
• External Media: While compatible with NAS drives, ensure your network connection is stable during indexing to prevent corruption of the .photo-index/ sidecar files. Always maintain a separate backup of your photo library before performing massive automation tasks.