The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads the luxin-cli package from the public npm registry and fetches skill definitions from the danielgwilson/luxin repository on GitHub.
[REMOTE_CODE_EXECUTION]: The instruction npx -y luxin-cli@latest downloads and executes code from a remote registry. The -y flag bypasses user confirmation, and the @latest tag ensures the most recent version is run without local verification or version pinning.
[COMMAND_EXECUTION]: The skill requires the execution of shell commands to set environment variables (export), create workspace directories (mkdir), and invoke the CLI tool.
[PROMPT_INJECTION]: The skill processes user-supplied strings which are passed to the luxin-cli for image generation. This creates an attack surface for indirect prompt injection into a tool with network capabilities.
Ingestion points: User-supplied strings in the --prompt field of the luxin-cli create command in SKILL.md.
Boundary markers: None identified; the prompt is passed directly as a command-line argument.
Capability inventory: The luxin-cli tool performs network requests to api.luxin.sh and interacts with the local file system.
Sanitization: No evidence of input sanitization or validation is present in the skill instructions.

image-skill