The Agent Skills Directory

[COMMAND_EXECUTION] (MEDIUM): The skill contains Python scripts that use subprocess.run to execute the nmap utility and Jenkins configurations that run shell commands like npm audit and snyk test.\n- [DATA_EXFILTRATION] (MEDIUM): Several scripts are designed to programmatically retrieve sensitive data from security platforms (Splunk, Nessus) and cloud environments (AWS S3). One script specifically includes logic to send automated reports via SMTP (smtplib), which could be misused to exfiltrate data if inputs or destinations are manipulated.\n- [PROMPT_INJECTION] (MEDIUM): The skill is vulnerable to indirect prompt injection (Category 8) because it ingests untrusted data from external sources, including Nmap scan results, Splunk logs, and S3 metadata. It lacks explicit boundary markers or sanitization logic to prevent malicious content in these tool outputs from influencing subsequent agent actions or decisions.\n- [CREDENTIALS_UNSAFE] (LOW): Multiple code snippets use hardcoded placeholder credentials (e.g., 'username', 'password', 'access_key:secret_key'). While intended as examples, they represent a risk of accidental deployment into production environments.

security-automation