← 返回 Skills 市场
SMTools Image Generation Skill
作者
Sergey Mikhaylov
· GitHub ↗
· v1.9.1
· MIT-0
335
总下载
0
收藏
1
当前安装
18
版本数
在 OpenClaw 中安装
/install smtools-image-generation
功能描述
Generate images from text prompts using AI models via OpenRouter, Kie.ai, or YandexART. Use when the user asks to generate, create, draw, or illustrate an im...
安全使用建议
This skill appears to do what it claims, but review these points before installing:
- Only provide API keys for providers you trust. OPENROUTER_API_KEY is required for default use; KIE and Yandex keys are optional and only used when you explicitly select those providers.
- Using the -i / edit option will read the given local file and send its contents to the external provider (base64 upload or provider-specific task). Do not pass sensitive local files unless you consent to uploading them.
- The provided setup.sh installs packages from PyPI and creates files under the skill directory (config.json, .env template, output/). Inspect setup.sh if you prefer to run in an isolated environment or a controlled virtualenv.
- There is a minor packaging/doc mismatch: check.sh looks for a 'dotenv' module but requirements.txt only lists 'requests' — expect the dependency check to warn about missing dotenv. You may need to install python-dotenv or adjust the scripts if you rely on .env parsing semantics.
- If you plan to enable autonomous invocation for this skill in an agent, be cautious about granting it the ability to generate or edit images that include content from local files or to use your provided API keys without supervision.
功能分析
Type: OpenClaw Skill
Name: smtools-image-generation
Version: 1.9.1
The skill is a well-structured image generation tool for OpenClaw, supporting multiple providers including OpenRouter, Kie.ai, and YandexART. The code follows a clean strategy pattern, implementing standard API interactions for image generation and editing. No evidence of data exfiltration, malicious execution, or harmful prompt injection was found; the scripts (setup.sh, check.sh, run.sh) and Python logic (generate.py, config_manager.py) are focused entirely on the stated functionality.
能力评估
Purpose & Capability
Name/description match the code and runtime behavior: the skill implements OpenRouter, Kie.ai, and YandexART providers and only requires python3 and provider API keys. The presence of provider implementations in scripts aligns with the declared purpose. Minor inconsistency: the SKILL.md frontmatter lists only KIE_API_KEY in optionalEnv but the skill also supports Yandex envs (YANDEX_IAM_TOKEN / YANDEX_FOLDER_ID) in code and docs.
Instruction Scope
Runtime instructions are scoped to generating or editing images. However, when using the editing mode (-i), the skill will read any local file path provided and embed/read it (OpenRouter encodes the file to base64 and sends it; Kie expects a URL and the code will include the provided value). That means local images are uploaded to external provider APIs — the SKILL.md does not explicitly warn users that passing a local path will upload the file. The config loader also reads a .env file and populates environment variables (no exfiltration in code), and all external calls go to the stated provider endpoints.
Install Mechanism
There is no platform install spec; installation is via provided setup.sh which installs dependencies via pip into a vendor directory and a venv (only requests is in requirements.txt). This is expected for a Python skill. Issues to note: scripts/check.sh expects a 'dotenv' import but requirements.txt does not include python-dotenv, so dependency checks may report missing 'dotenv' even after running setup.sh. setup.sh writes files under the skill directory (config.json, .env template, output/) which is normal but should be reviewed before running.
Credentials
PrimaryEnv is OPENROUTER_API_KEY which matches the default provider. Optional provider keys (KIE_API_KEY and YANDEX_IAM_TOKEN/YANDEX_FOLDER_ID) are used only for their respective providers. This is proportionate to the functionality. Minor mismatch: the SKILL.md frontmatter's optionalEnv only lists KIE_API_KEY and omits the Yandex env vars that the code requires if YandexART is selected.
Persistence & Privilege
The skill does not request always:true, does not modify other skills or system-wide settings, and only writes files inside its own skill directory (config.json, .env template, output images). Agent invocation settings are defaults (user-invocable and can run autonomously).
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install smtools-image-generation - 安装完成后，直接呼叫该 Skill 的名称或使用
/smtools-image-generation触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.9.1
- Add YandexART as a supported provider for image generation.
- Update description and provider section to include YandexART details.
- Specify required environment variables for YandexART: YANDEX_IAM_TOKEN and YANDEX_FOLDER_ID.
- Clarify instructions on when and how to use each provider.
v1.9.0
- Added support for YandexART as a new image generation provider.
- Updated provider selection and documentation to include YandexART, its models, and required environment variables.
- Modified configuration example and code to recognize the new provider and related settings.
- Maintained existing functionality for OpenRouter and Kie.ai providers.
v1.8.1
- Clarified instructions for using the skill to edit, modify, or transform existing images.
- Added a new "Image Editing" section explaining the use of the `-i` flag for input images, with decision guidance and prompt tips.
- No changes to commands or core functionality; documentation update only.
v1.8.0
**Image editing via input images is now supported.**
- Added support for image editing and transformation; you can now provide an input image for modification.
- New `-i, --input` option allows specifying an input image (path or URL) for editing.
- Documentation and examples updated with image editing instructions.
- Improved provider interfaces to handle image input where supported.
v1.7.2
- Added models `openai/gpt-image-1` and `nano-banana-2` to the supported list in SKILL.md.
- Updated the provider/model documentation for better clarity and completeness.
- No functional script changes; documentation improvements only.
v1.7.1
smtools-image-generation 1.7.1
- Updated README.md for improved documentation clarity; no functional or code changes.
v1.7.0
- Improved scripts/providers/openrouter_provider.py to enhance image generation reliability.
- No user-facing changes to documentation or usage instructions.
v1.5.0
- Added scripts/providers/openrouter_provider.py.
- Initial provider implementation for OpenRouter in the image generation pipeline.
v1.4.1
smtools-image-generation 1.4.1
- Internal: Modified setup.sh script.
- No user-facing feature or documentation changes.
v1.4.0
smtools-image-generation 1.4.0
- Added support for environment-based configuration via a new config manager.
- Improved image generation script for better provider handling and error management.
- Updated requirements.txt to include any new dependencies.
v1.3.2
- Updated setup.sh to improve setup or initialization (details not specified).
- No changes to usage, options, or functionality described in documentation.
- All existing features, providers, and error handling remain unchanged.
v1.3.1
- Fixed a bug in scripts/providers/kie_provider.py.
- No user-facing changes; internal code quality improvement.
v1.3.0
- Added KIE_API_KEY to the list of optional environment variables in metadata for improved provider flexibility.
- No changes to usage, output, or interface—only metadata was updated.
v1.2.3
- Added CLAUDE.md file with relevant information.
- Updated scripts/generate.py and setup.sh to improve functionality or address issues.
- No changes to user-facing usage or core features.
v1.2.2
smtools-image-generation 1.2.2
- Updated scripts/generate.py (details not specified).
- No changes to documentation or SKILL.md.
- No new features or breaking changes documented.
v1.2.1
**Changelog for smtools-image-generation v1.2.1**
- Added a `run.sh` script for simplified usage and updated all usage instructions to use `bash scripts/run.sh` instead of direct Python calls.
- Updated documentation and examples in SKILL.md and README.md to reflect the preferred use of the new `run.sh` script.
- Changed the default OpenRouter model example from `openai/gpt-image-1` to `google/gemini-3.1-flash-image-preview`.
- Minor documentation and usage improvements for clarity.
- No changes to image generation core logic.
v1.1.0
- Documentation updated: README.md revised for improved clarity and usage guidance.
- No changes to code or functionality in this version.
v1.0.0
Initial release of the smtools-image-generation skill.
- Generate images from text prompts using OpenRouter (default) or Kie.ai (optional, async).
- Command-line script with options for prompts, model selection, output path, and provider.
- Supports error handling for missing API keys, network issues, and timeouts.
- Includes setup instructions and security guidelines.
- Outputs JSON with image path and details upon successful image generation.
元数据
常见问题
SMTools Image Generation Skill 是什么？
Generate images from text prompts using AI models via OpenRouter, Kie.ai, or YandexART. Use when the user asks to generate, create, draw, or illustrate an im... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 335 次。
如何安装 SMTools Image Generation Skill？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install smtools-image-generation」即可一键安装，无需额外配置。
SMTools Image Generation Skill 是免费的吗？
是的，SMTools Image Generation Skill 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
SMTools Image Generation Skill 支持哪些平台？
SMTools Image Generation Skill 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 SMTools Image Generation Skill？
由 Sergey Mikhaylov（@bzsega）开发并维护，当前版本 v1.9.1。
推荐 Skills