The Agent Skills Directory

[Prompt Injection] (SAFE): No instructions attempting to override agent behavior or bypass safety filters were detected in the skill metadata or instructions.
[Data Exposure & Exfiltration] (SAFE): The skill performs local string manipulation only. No network calls (curl, wget, etc.) or sensitive file path access (~/.ssh, ~/.aws) were found.
[Indirect Prompt Injection] (LOW): The skill processes untrusted user input provided via command line arguments in scripts/formatter.py.
Ingestion points: User-provided text passed through sys.argv.
Boundary markers: None present.
Capability inventory: Limited to string manipulation (regex, capitalization). No file-writes, subprocess calls, or network operations.
Sanitization: None, but the lack of dangerous capabilities mitigates the risk.
[Remote Code Execution] (SAFE): No patterns for downloading or executing remote scripts (e.g., pipe to bash) were identified.
[Dynamic Execution] (SAFE): The Python scripts use standard library functions and do not employ eval(), exec(), or other dynamic code execution methods.

simple-formatter