The Agent Skills Directory

[SAFE]: The skill defines a set of commands for interacting with Google Search Console through a specialized CLI tool. All operations are documented with safety-first flags like --readonly and --dry-run to prevent accidental mutations.\n- [SAFE]: Includes explicit instructions for handling untrusted content using the --wrap-untrusted flag, which is a recommended practice to mitigate indirect prompt injection from external data sources. Evidence chain: (1) Ingestion points: Google Search Console via query/sitemaps/sites commands; (2) Boundary markers: Encourages use of --wrap-untrusted; (3) Capability inventory: gog CLI operations; (4) Sanitization: Implementation of content wrapping.\n- [SAFE]: No suspicious network activity, hardcoded credentials, or obfuscated code were detected.

gog-searchconsole