The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the gog command-line utility to interact with Google Contacts APIs for searching, listing, creating, and modifying contact records.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from Google Contacts which may contain attacker-controlled content designed to influence agent behavior.
Ingestion points: Untrusted data enters the agent's context through output from commands such as gog contacts search, list, get, directory, and other in SKILL.md.
Boundary markers: The skill documentation explicitly instructs the agent to use the --wrap-untrusted flag to delimit and identify potentially dangerous content in SKILL.md.
Capability inventory: The skill provides full CRUD capabilities (create, update, delete) and export functionality using the gog CLI tool in SKILL.md.
Sanitization: The instruction to use --wrap-untrusted serves as a sanitization and delimiting mechanism to help the agent distinguish between internal instructions and external data.

gog-contacts