The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes content from external websites (via domain and url parameters) and competitor pages. This ingestion of untrusted data creates a surface for indirect prompt injection.\n
Ingestion points: Website content scanned by /geo-content-audit and citation data from /geo-citation-monitor in skill.md.\n
Boundary markers: None detected; the skill lacks specific delimiters to separate untrusted data from instructions during processing.\n
Capability inventory: The skill utilizes Bash, Write, and Edit tools, and can perform network-based alerting (Email, Slack, SMS) as defined in the alert_rules.\n
Sanitization: No explicit sanitization or filtering of the external content is implemented before it is interpolated into reports or notifications.\n- [DATA_EXFILTRATION]: The skill is designed to send technical data and metrics to external channels, including email addresses (seo-manager@company.com), Slack channels (#seo-alerts), and SMS numbers. While this is the intended workflow, it provides a functional pathway for data to be moved out of the agent's immediate environment.\n- [COMMAND_EXECUTION]: The skill defines and orchestrates a complex workflow involving custom shell-like commands (e.g., /geo-content-audit, /generate-report) and utilizes the Bash tool to perform multi-step automated tasks and parallel processing.

automation-coordinator