The Agent Skills Directory

[Prompt Injection] (LOW): Surface for indirect prompt injection exists because the skill processes untrusted data from the Kubernetes cluster (e.g., logs and events).
Ingestion points: Data enters the agent context through scripts/discover_apis.py and various kubectl commands (logs, events, resource statuses) defined in references/operator-checks.md.
Boundary markers: Absent. The instructions do not provide delimiters or warnings to ignore instructions embedded within the cluster data.
Capability inventory: Includes read-only kubectl access and execution of local Python scripts via uv.
Sanitization: Absent. Logs and event messages are aggregated into reports without filtering or escaping content that might influence the LLM.
[External Downloads] (LOW): The skill depends on the official kubernetes Python package.
Dependency: kubernetes>=28.1.0 listed in scripts/discover_apis.py.
Source: Trusted organization (kubernetes). Per [TRUST-SCOPE-RULE], dependencies from trusted sources are categorized as LOW severity.

kubernetes-health