The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the gety CLI to perform searches, retrieve document content, and manage data connectors. This includes commands like gety search, gety doc, and gety connector.
[INDIRECT_PROMPT_INJECTION]: The skill processes content from local files, which represents a potential surface for indirect prompt injection if documents contain malicious instructions.
Ingestion points: Data is ingested via the gety doc command, which retrieves the full content of local documents (PDFs, docx, etc.).
Boundary markers: The instructions include best practices for handling large documents, such as redirecting output to temporary files and using grep or head to inspect specific sections, which reduces the direct exposure of the agent's context to untrusted data.
Capability inventory: The skill can read local files, execute shell commands via the CLI, and manage local data connectors.
Sanitization: The instructions emphasize synthesizing findings into natural language and citing sources, though they do not mandate specific sanitization of the document content itself.
[DYNAMIC_EXECUTION]: The gety connector install <directory> command allows for the installation of custom connectors from local directories, which involves loading and potentially executing logic within the Gety environment as part of its extensibility features.

gety-cli