The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes curl to perform cluster management and data retrieval from Elasticsearch indices.\n
Evidence: Multiple shell script examples in SKILL.md demonstrate using curl to update /_cluster/settings and perform /_search queries.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes and displays data from security audit logs which can include untrusted external input (e.g., usernames, IP addresses, or request bodies in failed login attempts).\n
Ingestion points: The Query Audit Events section in SKILL.md performs searches on .security-audit-* indices.\n
Boundary markers: Absent in the provided templates.\n
Capability inventory: The skill can modify cluster settings using curl -X PUT.\n
Sanitization: No sanitization or filtering of the queried log data is mentioned before presentation to the agent.

elasticsearch-audit