The Agent Skills Directory

[SAFE]: The skill consists of educational markdown documentation and Python audit scripts. No prompt injection, data exfiltration, or unauthorized command execution patterns were identified.
[COMMAND_EXECUTION]: The provided Python scripts (accessibility_checker.py and ux_audit.py) perform static analysis on local project files using regular expressions. These scripts do not execute the content of the files they analyze and do not perform any network operations.
[DATA_EXPOSURE & EXFILTRATION]: No sensitive file paths or hardcoded credentials were found. The scripts only read project-related file extensions (.html, .tsx, .css, etc.) and output results to the console.
[INDIRECT_PROMPT_INJECTION]: The audit scripts ingest untrusted data (source code from the project being audited). However, the risk is negligible as the scripts only perform pattern matching and do not use the data to influence agent behavior or execute dynamic commands.
Ingestion points: accessibility_checker.py and ux_audit.py read content from project files matching specific web extensions.
Boundary markers: Absent; the scripts read full file contents.
Capability inventory: File-read operations and console output via print() and json.dumps().
Sanitization: None; data is processed via regex for audit purposes.

frontend-design