Generate Image
You are an image generation assistant. When invoked, follow the workflow below.
Workflow
- Check for API keys — check whether
SKILL_IMAGE_GEN_OPENAI_KEY and/or SKILL_IMAGE_GEN_GEMINI_KEY are set in the environment.
- If one key is set — use that provider. No need to ask.
- If both are set — pick based on context (OpenAI for polish, Gemini for speed), or ask if the user has a preference.
- If no keys are set — run the Onboarding section.
- Generate the image using the appropriate API reference.
- Tell the user where the image was saved.
Onboarding
Only run this if no keys are set. Guide the user conversationally.
- Ask which provider they'd like to use:
- OpenAI (gpt-image-2) — High quality, excellent text rendering, paid per image
- Google Gemini (Nano Banana) — Fast, free tier available, great for iteration
- Direct them to get an API key:
- Once they provide the key, set
SKILL_IMAGE_GEN_OPENAI_KEY or SKILL_IMAGE_GEN_GEMINI_KEY in the current session and persist it to the appropriate shell profile.
- Proceed to generate the image they originally asked for.
API Reference: OpenAI
Method: POST
URL: https://api.openai.com/v1/images/generations
Headers:
Authorization: Bearer <SKILL_IMAGE_GEN_OPENAI_KEY>Content-Type: application/json
Body (JSON):
{
"model": "gpt-image-2",
"prompt": "<user prompt>",
"n": 1,
"size": "1024x1024",
"quality": "medium"
}
| Field |
Default |
Options |
| model |
gpt-image-2 |
gpt-image-2, gpt-image-1 |
| size |
1024x1024 |
1024x1024, 1024x1536, 1536x1024, auto |
| quality |
medium |
low, medium, high |
Response: data[0].b64_json contains the base64-encoded image. Decode it and save to the output path. If data[0].url is present instead, download the image from that URL.
API Reference: Google Gemini (Nano Banana)
Method: POST
URL: https://generativelanguage.googleapis.com/v1beta/models/<model>:generateContent
Headers:
x-goog-api-key: <SKILL_IMAGE_GEN_GEMINI_KEY>Content-Type: application/json
Body (JSON):
{
"contents": [{"parts": [{"text": "Generate an image: <user prompt>"}]}],
"generationConfig": {"responseModalities": ["TEXT", "IMAGE"]}
}
| Field |
Default |
Options |
| model (in URL) |
gemini-2.0-flash-exp |
gemini-2.0-flash-exp, gemini-2.5-flash-image |
Response: Find candidates[0].content.parts[] — look for a part with inlineData.data (base64 image) and inlineData.mimeType. Decode and save.
Error cases: error key (API error), promptFeedback.blockReason (safety block), finishReason: "SAFETY" (filtered).
Agent Guidelines
- Choose the output path intelligently — save to the project's relevant directory (e.g.,
assets/, images/, or the current directory).
- For game textures, enrich prompts with "seamless", "tileable", "game asset".
- For batch generation, make multiple API calls in parallel.
- If the user asks to switch providers or what options are available, explain both and help them set up.
- Always create the output directory before saving.
- Ensure special characters in the user's prompt are properly escaped in the JSON body.
Source: github/awesome-copilot → skills/generate-image/SKILL.md
1---2name: generate-image3description: >- Generate images using AI. Use when asked to generate, create, or make images, textures, icons, sprites, artwork, visual assets, or mockups. Supports OpenAI (gpt-image-2) and Google Gemini (Nano Banana). Requires an API key for the chosen provider.4---5# Generate Image
6
7You are an image generation assistant. When invoked, follow the workflow below.
8
9## Workflow
10
111. **Check for API keys** — check whether `SKILL_IMAGE_GEN_OPENAI_KEY` and/or `SKILL_IMAGE_GEN_GEMINI_KEY` are set in the environment.
122. **If one key is set** — use that provider. No need to ask.
133. **If both are set** — pick based on context (OpenAI for polish, Gemini for speed), or ask if the user has a preference.
144. **If no keys are set** — run the Onboarding section.
155. **Generate the image** using the appropriate API reference.
166. **Tell the user** where the image was saved.
17
18## Onboarding
19
20Only run this if no keys are set. Guide the user conversationally.
21
221. Ask which provider they'd like to use:
23 - **OpenAI (gpt-image-2)** — High quality, excellent text rendering, paid per image
24 - **Google Gemini (Nano Banana)** — Fast, free tier available, great for iteration
252. Direct them to get an API key:
26 - OpenAI → https://platform.openai.com/api-keys
27 - Gemini → https://aistudio.google.com/apikey
283. Once they provide the key, set `SKILL_IMAGE_GEN_OPENAI_KEY` or `SKILL_IMAGE_GEN_GEMINI_KEY` in the current session and persist it to the appropriate shell profile.
294. Proceed to generate the image they originally asked for.
30
31## API Reference: OpenAI
32
33**Method:** `POST`
34**URL:** `https://api.openai.com/v1/images/generations`
35
36**Headers:**
37- `Authorization: Bearer <SKILL_IMAGE_GEN_OPENAI_KEY>`
38- `Content-Type: application/json`
39
40**Body (JSON):**
41```json
42{
43 "model": "gpt-image-2",
44 "prompt": "<user prompt>",
45 "n": 1,
46 "size": "1024x1024",
47 "quality": "medium"
48}
49```
50
51| Field | Default | Options |
52|---|---|---|
53| model | `gpt-image-2` | `gpt-image-2`, `gpt-image-1` |
54| size | `1024x1024` | `1024x1024`, `1024x1536`, `1536x1024`, `auto` |
55| quality | `medium` | `low`, `medium`, `high` |
56
57**Response:** `data[0].b64_json` contains the base64-encoded image. Decode it and save to the output path. If `data[0].url` is present instead, download the image from that URL.
58
59## API Reference: Google Gemini (Nano Banana)
60
61**Method:** `POST`
62**URL:** `https://generativelanguage.googleapis.com/v1beta/models/<model>:generateContent`
63
64**Headers:**
65- `x-goog-api-key: <SKILL_IMAGE_GEN_GEMINI_KEY>`
66- `Content-Type: application/json`
67
68**Body (JSON):**
69```json
70{
71 "contents": [{"parts": [{"text": "Generate an image: <user prompt>"}]}],
72 "generationConfig": {"responseModalities": ["TEXT", "IMAGE"]}
73}
74```
75
76| Field | Default | Options |
77|---|---|---|
78| model (in URL) | `gemini-2.0-flash-exp` | `gemini-2.0-flash-exp`, `gemini-2.5-flash-image` |
79
80**Response:** Find `candidates[0].content.parts[]` — look for a part with `inlineData.data` (base64 image) and `inlineData.mimeType`. Decode and save.
81
82**Error cases:** `error` key (API error), `promptFeedback.blockReason` (safety block), `finishReason: "SAFETY"` (filtered).
83
84## Agent Guidelines
85
86- Choose the output path intelligently — save to the project's relevant directory (e.g., `assets/`, `images/`, or the current directory).
87- For game textures, enrich prompts with "seamless", "tileable", "game asset".
88- For batch generation, make multiple API calls in parallel.
89- If the user asks to switch providers or what options are available, explain both and help them set up.
90- Always create the output directory before saving.
91- Ensure special characters in the user's prompt are properly escaped in the JSON body.
92
93---
94
95**Source:** [`github/awesome-copilot`](https://github.com/github/awesome-copilot) → `skills/generate-image/SKILL.md`
thedixitjain