The Agent Skills Directory

[SAFE] (SAFE): No security issues detected in the script logic.
The script serves as a passive auditor for project architecture, using standard tools like grep and find to check for the presence of specific libraries (TanStack Query, Zustand) and patterns (prop drilling, type safety).
It utilizes safe bash practices, such as set -euo pipefail for error handling and -- in the cd command to prevent positional argument injection.
It lacks any destructive capabilities (no rm, mv, or file-writing operations outside of variable assignments).
No network-facing commands (curl, wget) or external data exfiltration vectors were found.
Although it processes untrusted data (project files), the script only reports findings via standard output and does not execute the content it finds, presenting a negligible risk of indirect prompt injection.

state-management