The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Automated security scanners have flagged the documentation URL https://jeffallan.github.io/claude-skills/skills/infrastructure/kubernetes-specialist/ as malicious (Blacklisted). Additionally, the main skill file SKILL.md has been flagged for negative reputation (FileRepMalware).
[REMOTE_CODE_EXECUTION]: The skill provides instructions for downloading and executing code from external sources, including piping remote scripts to a shell for tool installations (e.g., Istio and Linkerd setup scripts).
[CREDENTIALS_UNSAFE]: Multiple reference files, such as references/configuration.md, contain hardcoded credentials and secrets within example configurations. These include database passwords, API keys, and JWT secrets which could be inadvertently adopted into production environments.
[PRIVILEGE_ESCALATION]: The skill encourages high-privilege operations, such as using sudo to move binaries to system paths and defining container workloads with privileged: true security contexts.
[COMMAND_EXECUTION]: The skill relies on the execution of powerful shell commands and kubectl operations to manage cluster resources, which can be misused if the input is not strictly controlled.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests user-provided workload descriptions and security requirements to generate and potentially execute Kubernetes manifests.
Ingestion points: Workload definitions and security requirements processed by the agent in SKILL.md.
Boundary markers: Absent; no instructions are provided to the agent to treat embedded data as untrusted or to ignore instructions within the input.
Capability inventory: Includes arbitrary command execution via Jobs/CronJobs, network configuration via Ingress, and persistent storage access via PVCs.
Sanitization: No input validation or sanitization routines are implemented in the skill instructions.

kubernetes-specialist