The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The installation instructions require the user to clone a third-party repository and install dependencies from it.
Evidence: git clone https://github.com/drolosoft/immich-photo-manager.git and pip3 install -r src/requirements.txt in SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from an Immich instance, which could contain malicious instructions embedded in metadata or text found via OCR.
Ingestion points: The skill reads asset metadata, GPS data, and text extracted via OCR from the photo library (mentioned in SKILL.md).
Boundary markers: No specific delimiters or safety warnings for embedded instructions are mentioned in the documentation.
Capability inventory: The skill possesses the ability to move assets to trash, permanently delete assets (on request), merge/rename people, and write metadata/albums to the Immich instance.
Sanitization: The skill does not describe any sanitization or filtering of the external metadata or OCR content before it is processed by the agent.
[COMMAND_EXECUTION]: The installation and setup process involves executing multiple shell commands, including plugin installation and credential configuration.
Evidence: claude plugin marketplace add ./, claude plugin install immich-photo-manager, and /setup-immich-photo-manager command in SKILL.md.

immich-photo-manager