The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches documentation indices and configuration manifests from the vendor's official domain (openchoreo.dev) and GitHub repositories (github.com/openchoreo).
[COMMAND_EXECUTION]: The skill makes extensive use of system tools including kubectl, helm, and git to manage Kubernetes cluster states, install platform components, and clone source code.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to fetch and process remote data which the agent then follows as instructions, creating a potential attack surface.
Ingestion points: Documentation pages are retrieved via scripts/fetch-page.sh and repository contents are fetched via git clone from the vendor's GitHub organization.
Boundary markers: The skill lacks explicit delimiters or warnings to the agent to ignore potentially malicious instructions embedded within the fetched remote Markdown content.
Capability inventory: The agent has permissions to execute kubectl, helm, and general shell commands, which provide significant control over the target environment.
Sanitization: The skill performs basic validation of version strings and regex-escapes title strings used in searches, but the resulting remote content is treated as a trusted source of instructions.
[DYNAMIC_EXECUTION]: Installation playbooks and specific command sequences are dynamically resolved at runtime from remote vendor resources rather than being statically defined within the skill.

openchoreo-setup