The Agent Skills Directory

[COMMAND_EXECUTION]: The script generate-image.ts is vulnerable to path traversal. The --dir and --output arguments are used in path.join and path.resolve without sanitization, allowing a user-provided prompt to potentially trigger the writing or overwriting of sensitive files outside the intended directory (e.g., ~/.bashrc).
[EXTERNAL_DOWNLOADS]: The skill uses npx tsx to execute its logic. This triggers the download and execution of the tsx package and its dependencies from the npm registry at runtime.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data and has significant system capabilities.
Ingestion points: The prompt argument in generate-image.ts accepts arbitrary user-controlled text.
Boundary markers: None. The prompt is interpolated directly into a JSON request body for an external API.
Capability inventory: The script has file system write access (writeFileSync), directory creation (mkdirSync), and network access via the https module.
Sanitization: There is no sanitization or validation performed on the user-provided prompt or the resulting output paths.

generate-image