The Agent Skills Directory

[PROMPT_INJECTION] (SAFE): No instructions attempting to override agent behavior or safety filters were found in the skill markdown or documentation.
[DATA_EXFILTRATION] (SAFE): No hardcoded credentials, sensitive file path access, or network operations were detected. The scripts (anomaly_detection.py, similarity_metrics.py) operate strictly on data structures using standard libraries.
[REMOTE_CODE_EXECUTION] (SAFE): No patterns of remote script execution, dynamic code evaluation (eval/exec), or untrusted data deserialization were identified.
[EXTERNAL_DOWNLOADS] (SAFE): Dependencies listed in requirements.txt (pandas, numpy) are standard, well-known libraries from trusted sources.
[COMMAND_EXECUTION] (SAFE): No use of subprocess, os.system, or similar commands to execute shell instructions was found.
[INDIRECT_PROMPT_INJECTION] (SAFE): While the skill processes external data (datasets), it lacks exploitable capabilities like network access or system command execution, minimizing the risk surface. Ingestion points include dataframes passed to analytical functions; however, no sanitization is present as no dangerous downstream operations are performed.

data-quality