The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon data from external sources such as incident alerts, Slack threads, and monitoring dashboards. Malicious instructions embedded in these external sources could influence the agent's actions, particularly when following the provided runbook templates for infrastructure management.
Ingestion points: Slack incident channels, monitoring alerts, and user reports (referenced in the 'Incident Detection' workflow).
Boundary markers: No specific delimiters or instructions to ignore embedded content are provided to isolate untrusted external content.
Capability inventory: The skill provides templates for high-privilege kubectl commands (get, rollout, restart, scale) within the 'Runbook Template' section.
Sanitization: No explicit sanitization or validation of incoming incident data is defined in the instructions.

Incident Response Commander