The Agent Skills Directory

[DATA_EXFILTRATION]: The skill transmits the user-provided LNAPI_KEY and prompt content to a third-party proxy endpoint at https://lnapi.com. While this is the intended purpose of the skill, users should be aware that their credentials and data are shared with this specific external service.
[COMMAND_EXECUTION]: The script scripts/main.ts performs file system operations, including recursive directory creation (mkdir) and file writing (writeFile), based on user-controlled CLI arguments for output paths. This allows the agent to write files to the local system.
[PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection due to how it processes input data.
Ingestion points: Prompts are ingested via CLI arguments, concatenated from local files using the --promptfiles flag, or read from standard input in scripts/main.ts.
Boundary markers: Absent; prompt strings are interpolated directly into the API request body without delimiters or instructions to ignore embedded commands.
Capability inventory: The skill has network access (via fetch in scripts/providers/google.ts) and file system write access (in scripts/main.ts).
Sanitization: No sanitization, filtering, or validation is performed on the prompt content before it is sent to the image generation model.

banana-proxy