The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to automatically execute the lzc-cli box default command to retrieve environment configuration. This execution is performed without requesting user confirmation or providing the user an opportunity to review the action.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to take user requirements and generate complex deployment configurations, including docker-compose.yml, lzc-manifest.yml, and Caddy server configuration scripts. It lacks instructions for sanitizing user-provided strings or using boundary markers, which could allow a malicious user to inject commands or unintended configurations into the target environment through the generated setup_script or volume mappings.
Ingestion points: User prompts requesting the construction, modification, or packaging of AI application configurations as defined in SKILL.md.
Boundary markers: No delimiters or 'ignore embedded instructions' warnings are mandated for the agent to use when interpolating user data into deployment templates.
Capability inventory: The agent generates shell script blocks (e.g., Caddy setup_script), writes docker-compose.yml service definitions, and executes the lzc-cli tool.
Sanitization: The skill provides no guidance for escaping, validating, or filtering user-supplied values before they are integrated into generated scripts or YAML configuration files.
[EXTERNAL_DOWNLOADS]: The skill provides examples that reference container images from registry.lazycat.cloud. These are specific vendor resources for the 'Lazy Cat' platform described in the documentation and are used for standard deployment purposes.

lazycat-aipod-developer