The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script scripts/check_profile.py to validate the user's branding profile (person-dna.md). Technical analysis of the script confirms it uses only standard libraries (argparse, re, datetime, pathlib) to perform structural checks and does not perform any network operations, persistent changes, or arbitrary command execution.
[SAFE]: The skill handles user-provided branding data (person-dna.md) according to its stated purpose. The instructions include explicit safeguards to prevent the agent from scanning unrelated private directories or leaking the content of the profile. All operations are performed locally within the agent's context.

yetong-cover