The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were detected. The skill's behavior is consistent with its stated purpose of production troubleshooting.
[COMMAND_EXECUTION]: The skill utilizes kubectl and a custom tool agent-tools-k8s to interact with Kubernetes clusters. These tools are used for diagnostic operations such as fetching logs, checking resource usage (top), and inspecting pod configurations. High-privilege commands like kubectl exec and kubectl rollout restart are documented for specific troubleshooting scenarios (e.g., DNS testing, database connection counts, and service recovery), which is appropriate for an SRE/DevOps context.
[DATA_EXPOSURE]: The skill facilitates access to application logs and environment configurations (Helm values). While these files may contain sensitive environment data or PII, the skill does not contain any patterns for exfiltrating this data to external or untrusted domains. The use of Sentry for trace analysis is a standard industry practice.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from application logs and Sentry traces. This creates a potential surface for indirect prompt injection if malicious payloads are logged by an application. However, the skill's instructions focus on keyword searching (e.g., ERROR, [DB Pool]) and manual inspection rather than automated execution of content found within the logs.

production-troubleshooting