nano-banana、gpt-image

Generate images via VAPI's OpenAI-compatible Images API. Supports nano-banana and gpt-image model series. Default model is nano-banana-pro. Images are NOT sa...

Security Scan

The skill's code, instructions, and requested environment variables line up with its stated purpose (calling a VAPI-compatible image generation API); there are no unexplained credentials, installs, or hidden endpoints.

Purpose & Capability

Name/description (image generation via VAPI) match the actual behavior: the script calls {base_url}/images/generations and /images/edits and consumes VAPI_API_KEY and VAPI_BASE_URL. No unrelated credentials or binaries are requested.

Instruction Scope

SKILL.md tells the agent to run the included Python script with prompt/model/save flags. The script only reads the two declared env vars, may download provided input URLs, calls only the configured base_url endpoints, and writes files under ~/.openclaw. There are no instructions to read unrelated system files or exfiltrate other data.

Install Mechanism

No install spec (instruction-only) and the included Python script uses only stdlib modules and urllib; nothing is downloaded from arbitrary URLs during installation. Risk from install mechanism is low.

Credentials

Only VAPI_API_KEY (primary) and VAPI_BASE_URL are required, which are appropriate for an API client. No unrelated secrets or system config paths are requested.

Persistence & Privilege

Skill is not always:true and does not request elevated or persistent platform privileges. It will create and write to ~/.openclaw media/ and oss/ directories as documented (expected behavior for save flags).

Assessment

This skill appears to do exactly what it says: it sends your VAPI_API_KEY to whatever VAPI_BASE_URL you configure and will download/save images to ~/.openclaw by default when requested. Before installing, confirm you trust the VAPI provider (default base URL is https://api.v3.cm/v1). Be aware that: the script will download any input image URL you pass and may download image URLs returned by the API, it will create ~/.openclaw/media (or ~/.openclaw/oss) when saving, and gpt-image models always save files. If you do not trust the configured base URL or do not want files written under your home directory, do not install or run this skill.

VAPI Image Gen

Generate images via VAPI's OpenAI-compatible Images API (/images/generations).

Basic Usage

python3 {baseDir}/scripts/gen.py --prompt "your prompt here"

Options

# Specify model (default: nano-banana-pro)
python3 {baseDir}/scripts/gen.py --prompt "..." --model nano-banana-pro

# Higher resolution (append -2k or -4k to model name)
python3 {baseDir}/scripts/gen.py --prompt "..." --model nano-banana-pro-2k

# Save image locally (default: NO save, URL only)
python3 {baseDir}/scripts/gen.py --prompt "..." --save

# Save to OSS directory (~/.openclaw/oss/)
python3 {baseDir}/scripts/gen.py --prompt "..." --oss

# Save to custom directory
python3 {baseDir}/scripts/gen.py --prompt "..." --save --out-dir /path/to/dir

# Aspect ratio
python3 {baseDir}/scripts/gen.py --prompt "..." --aspect-ratio 16:9

# Image count
python3 {baseDir}/scripts/gen.py --prompt "..." --count 2

Supported Models

nano-banana series (returns URL, fast):

• nano-banana
• nano-banana-pro ← default
• nano-banana-2
• nano-banana-pro-2k / nano-banana-pro-4k (high res)

gpt-image series (returns base64, always saved):

• gpt-image-1
• gpt-image-1.5

Config

Set in .env or ~/.openclaw/openclaw.json:

• VAPI_API_KEY — your VAPI API key (required)
• VAPI_BASE_URL — base URL with /v1 suffix (required); defaults to https://api.v3.cm/v1 if not set

Save Behavior

• Default: no local file, output MEDIA:<url> directly
• --save: save to ~/.openclaw/media/
• --oss: save to ~/.openclaw/oss/
• gpt-image models always save to media/ (API returns base64 only)