The Agent Skills Directory

[SAFE]: No malicious patterns detected. The skill uses declarative YAML and Markdown to define security logic, monitoring rules, and response workflows. All external references are conceptual or template-based placeholders.
[PROMPT_INJECTION]: Potential surface for indirect prompt injection via untrusted data ingestion. 1. Ingestion points: Security logs, incident data, alert rules, and content from phishing analysis steps in SKILL.md. 2. Boundary markers: Absent. The skill uses template placeholders like {{event_context}} to interpolate data into alerts without explicit delimiters or instructions to ignore potential commands within that data. 3. Capability inventory: Includes siem_query, alert_create, incident_manage, and compliance_check tools. 4. Sanitization: No explicit sanitization or validation of input data is described within the skill body.
[NO_CODE]: The skill consists exclusively of documentation and configuration in Markdown/YAML format and contains no executable scripts or external package dependencies.

Security Monitoring