The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. It ingests untrusted data from user requests and interpolates them into sensitive command contexts.\n
Ingestion points: Request parameters for volume management, backup targets, and disaster recovery configuration in SKILL.md.\n
Boundary markers: No explicit delimiters or instructions are used to separate user data from control instructions in the generated scripts or manifests.\n
Capability inventory: The skill leverages helm, kubectl, and bash to interact with the Kubernetes cluster and external networks.\n
Sanitization: There is no evidence of input validation or escaping for parameters such as $volume, $backup_target, or $pvc_name.\n- [COMMAND_EXECUTION]: The skill generates and executes shell commands for cluster management and performance tuning. While expected for a DevOps skill, the inclusion of system-level commands like sysctl and ip link in the documentation requires elevated privileges.\n- [EXTERNAL_DOWNLOADS]: The skill fetches Helm charts and container images from the official Longhorn project (charts.longhorn.io, longhornio/longhorn-engine). These are well-known services associated with the CNCF Longhorn project.

longhorn