The Agent Skills Directory

[SAFE]: No malicious patterns detected across all evaluated threat categories.
[COMMAND_EXECUTION]: The script scripts/examine_region.py performs standard image processing (cropping and contrast enhancement) using the Pillow library. It uses a safe argument parsing approach and does not invoke shell commands or interact with sensitive system files.
[DATA_EXFILTRATION]: The skill uses WebFetch and WebSearch for legitimate geolocation verification. No logic exists to exfiltrate user data or credentials to external domains.
[PROMPT_INJECTION]: The instructions focus purely on geolocation methodology and do not contain attempts to override agent safety guidelines or system prompts.
[EXTERNAL_DOWNLOADS]: No external scripts or packages are downloaded at runtime.

geolocation-skill