The Agent Skills Directory

[Prompt Injection] (SAFE): No instructions found that attempt to override agent behavior, bypass safety filters, or extract system prompts.
[Data Exposure & Exfiltration] (SAFE): No hardcoded credentials or access to sensitive local file paths (e.g., SSH keys, environment files) detected. No network operations were found in the code snippets.
[Obfuscation] (SAFE): The content is clear and readable with no Base64, zero-width characters, or hidden Unicode tags.
[Unverifiable Dependencies & Remote Code Execution] (SAFE): References standard, well-known machine learning libraries (shap, xgboost, numpy, etc.). There are no commands to download or execute remote scripts.
[Privilege Escalation] (SAFE): No administrative commands or system modification operations (e.g., sudo, chmod) are present.
[Persistence Mechanisms] (SAFE): No attempts to modify startup scripts, shell profiles, or scheduled tasks.
[Metadata Poisoning] (SAFE): Metadata accurately reflects the skill's purpose for model explainability.
[Indirect Prompt Injection] (LOW): While the skill involves processing data (model inputs/outputs), it does not display any vulnerability to prompt injection through data ingestion. It functions as a standard analytical tool.
[Time-Delayed / Conditional Attacks] (SAFE): No logic exists that triggers behavior based on time, date, or environment variables.
[Dynamic Execution] (SAFE): No use of eval(), exec(), or runtime compilation of generated code.

shap