The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes node:child_process to execute a vendor-specific command-line tool (dlazy) to perform image generation tasks, as documented in the skill's reference guides.
[EXTERNAL_DOWNLOADS]: The skill performs network requests to well-known AI services (including OpenAI, Google Gemini, and Replicate) and the vendor's official infrastructure (dlazy.com) to facilitate image generation and analysis.
[DATA_EXFILTRATION]: The skill reads local image files and configuration data (such as brand.yaml) to transmit them to authorized AI provider endpoints for processing, which is consistent with its primary purpose of image compliance optimization.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection due to the ingestion of untrusted data.
Ingestion points: User-provided prompts and image paths in scripts/gen.mjs, and brand settings in brand.yaml via scripts/brand.mjs.
Boundary markers: The skill does not employ explicit delimiters when appending brand constraints to the generation prompt.
Capability inventory: The skill can execute local binaries via spawn, perform network operations via fetch, and perform file system writes.
Sanitization: No specific prompt filtering or sanitization was observed before interpolation into AI requests.

platform-compliance