Grok Imagine

Generate images and videos using xAI Grok Imagine Extended. Text-to-image, image editing, text-to-video, image-to-video. Use when: user asks to generate, cre...

Security Scan

The skill's code, declared requirements, and runtime instructions are consistent with an xAI Grok Imagine image/video generator and do not request unrelated credentials or surprising system access.

Purpose & Capability

Name/description, required XAI_API_KEY, endpoints (api.x.ai), and the provided script all align with an image/video generation skill. Required capabilities are proportional to the stated purpose.

Instruction Scope

Runtime instructions tell the agent to run the included script and only reference XAI_API_KEY, a local keys.txt fallback, and output paths. The script does network calls to the xAI API, downloads returned media to disk, and polls for video status as documented — all within the scope. Minor mismatch: SKILL.md also claims the key can be read from ~/.openclaw/openclaw.json, but the included script only implements env var, explicit --api-key, and ~/keys.txt lookup.

Install Mechanism

No install spec; the skill is instruction-only with a bundled script. Nothing is downloaded or executed at install time.

Credentials

Only XAI_API_KEY is required (declared as primary). The script optionally reads ~/keys.txt for a fallback key (documented). There are no unrelated secrets requested. Users should note the script will read keys.txt if present — ensure that file's contents and permissions are acceptable.

Persistence & Privilege

always is false and the skill does not request persistent/privileged presence or modify other skills or system-wide settings. Autonomous invocation is allowed but is the platform default.

Assessment

This skill appears to do what it says: call xAI's image/video endpoints, download media, and save files. Before installing, ensure you: (1) only provide an XAI_API_KEY you trust for image generation and monitor usage/costs (video polling can incur charges while it runs); (2) are comfortable with the script reading ~/keys.txt if that file exists (remove or secure it if not); (3) understand the skill will download remote media and write files to the provided output paths; and (4) note the SKILL.md mentions reading ~/.openclaw/openclaw.json but the bundled script does not — if you rely on that behavior, verify it or set the env/--api-key explicitly. Overall the package is coherent with its stated purpose.

Grok Imagine Extended (xAI Image & Video Generation)

Generate images and videos from text prompts using the xAI API.

Image Generation

python3 {baseDir}/scripts/generate_image.py --prompt "your image description" --filename "output.png"

With options:

python3 {baseDir}/scripts/generate_image.py --prompt "a cyberpunk city at night" --filename "city.png" --resolution 2k --aspect-ratio 16:9

Image Editing

Single source image:

python3 {baseDir}/scripts/generate_image.py --prompt "make it a watercolor painting" --filename "edited.png" -i "/path/to/source.jpg"

Multiple source images (up to 3):

python3 {baseDir}/scripts/generate_image.py --prompt "combine into one scene" --filename "combined.png" -i img1.png -i img2.png

Video Generation

Text-to-video:

python3 {baseDir}/scripts/generate_image.py --prompt "a cat walking through flowers" --filename "cat.mp4" --video --duration 5

Image-to-video (animate a still):

python3 {baseDir}/scripts/generate_image.py --prompt "add gentle camera zoom and wind" --filename "animated.mp4" --video -i photo.jpg --duration 5

Models

Model	Type	Cost
grok-imagine-image	Image (default)	$0.02/img
grok-imagine-image-pro	Image (high quality)	$0.07/img
grok-imagine-video	Video (auto for --video)	$0.05/sec

Select model with --model grok-imagine-image-pro. Video mode always uses grok-imagine-video.

All Options

Flag	Description
--prompt, -p	Text description (required)
--filename, -f	Output path (required)
-i	Input image for editing/animation (repeatable, max 3 for images, 1 for video)
--model, -m	Image model (default: grok-imagine-image)
--aspect-ratio, -a	1:1, 16:9, 9:16, 4:3, 3:4, etc.
--resolution, -r	Image: 1k/2k. Video: 480p/720p
--n	Number of images 1-10 (default 1)
--video	Generate video instead of image
--duration, -d	Video duration 1-15 seconds (default 5)
--api-key, -k	Override XAI_API_KEY

API Key

• XAI_API_KEY env var
• Or set skills."grok-imagine".apiKey / skills."grok-imagine".env.XAI_API_KEY in ~/.openclaw/openclaw.json
• Or auto-read from ~/keys.txt

Notes

• Use timestamps in filenames: 2026-03-01-cyberpunk-city.png
• The script prints a MEDIA: line for OpenClaw to auto-attach on supported chat providers
• Do not read the image back; report the saved path only
• Image URLs from xAI are temporary; the script downloads them immediately
• Video generation is async and polls until done (can take 1-5 minutes)
• 2k resolution returns PNG; 1k returns JPEG