The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from Google Photos (such as album titles, descriptions, and media item metadata) which could be used to deliver indirect prompt injection attacks.
Ingestion points: Data is ingested through actions such as googlephotos_list_albums, googlephotos_list_media_items, and googlephotos_search_media_items in SKILL.md.
Boundary markers: There are no explicit instructions or delimiters provided to the agent to ignore potentially malicious instructions embedded in the photo metadata.
Capability inventory: The skill enables the agent to execute actions (reads and writes) via the npx @useclawlink/cli run command based on the processed data.
Sanitization: No evidence of sanitization or validation of the external content is present in the skill instructions.
[COMMAND_EXECUTION]: The skill relies on the execution of shell commands using the npx @useclawlink/cli tool for all operations, including authentication, discovery, and execution of actions.
[EXTERNAL_DOWNLOADS]: The skill instructions utilize npx to download and execute the @useclawlink/cli package from the public NPM registry. This is the official tool provided by the vendor (ClawLink) for the service integrated by this skill.

googlephotos