AI Image Generation

This skill appears coherent for multi‑provider image generation. Before installing: - Understand that prompts and any reference images you send will be transmitted to whichever provider you choose (OpenAI, Google, FLUX/BFL, Replicate, Leonardo, Ideogram, Midjourney via Discord). Only use providers you trust for sensitive content. - The skill stores local preferences and optional history under ~/image-generation/. If you want no local persistence, don't create or populate memory.md/history.md. - To use provider APIs you must set the provider API keys in your environment (the skill lists these as optional metadata). It will only check for presence (test -n) and the docs explicitly say it does not store API keys—still, avoid pasting secrets into chat. - Webhook examples exist (Replicate, etc.). Only configure webhook callbacks to servers you control; otherwise prefer polling to avoid sending data to third parties. - Because this is instruction-only (no install), no code is downloaded or installed by the skill itself — network calls happen only when you instruct the agent to call a provider. If privacy or offline-only operation is required, prefer the local routes (Stable Diffusion / flux-schnell) described in the docs and avoid setting cloud provider keys.