← 返回 Skills 市场
Jimeng AI Image Generator
作者
luis1213899
· GitHub ↗
· v1.0.2
· MIT-0
65
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install xingxiaolu-jimeng
功能描述
即梦AI生图与视频生成工具。通过火山引擎API调用即梦AI，支持文生图、图片生成、视频生成。内置50+提示词模板覆盖电商、社交、商业设计、视频等场景。
安全使用建议
This skill otherwise behaves like a normal Volcengine/Jimeng client, but it will automatically read credentials from ~/.openclaw/workspace/secrets.json under the key name 'jimeng' (AccessKeyId + SecretAccessKey) even though the registry doesn't declare any required credentials. Before installing or running it, verify: (1) the 'jimeng' entry in your token-manager/secrets.json is intended for this tool and has minimal permissions; (2) you trust the skill source (owner/homepage not provided); (3) inspect the secrets.json file and the scripts yourself if possible. If you don't want the skill to access your saved keys, do not install/run it or move/remove the 'jimeng' entry. Also note the skill will write image/video files to the agent workspace (temp_output.png, output.png).
功能分析
Type: OpenClaw Skill
Name: xingxiaolu-jimeng
Version: 1.0.2
The skill bundle provides a legitimate interface for Jimeng AI (Volcengine) image and video generation. It includes scripts (generate.py, video.py, query.py) for task submission, polling, and result retrieval, using standard HMAC-SHA256 signing for API authentication against the official Volcengine endpoint (visual.volcengineapi.com). Credentials are appropriately retrieved from the local OpenClaw secrets store (~/.openclaw/workspace/secrets.json). The code is well-structured, lacks any indicators of malicious intent, and the SKILL.md instructions are purely functional and helpful for the agent's operation.
能力评估
Purpose & Capability
Name/description and the included Python scripts line up: they call Volcengine's visual.volcengineapi.com to submit image/video generation tasks. Requiring python is expected. However, the skill uses an on-disk secrets entry named 'jimeng' (AccessKeyId + SecretAccessKey) in ~/.openclaw/workspace/secrets.json; the registry metadata declares no required credentials or config paths. This is an incoherence: a networked cloud API key is required but not declared.
Instruction Scope
SKILL.md explicitly instructs the agent to use the token-manager 'jimeng' key and to run the included scripts. The scripts read ~/.openclaw/workspace/secrets.json, sign requests, call Volcengine, poll for results, and save image/video outputs to local files (temp_output.png, output.png). Those operations are within the stated purpose, but they access local secrets that were not declared and write files under the skill/agent workspace — the instructions therefore require access to sensitive local data without that being surfaced in the skill metadata.
Install Mechanism
No install spec — instruction- and script-based. That minimizes supply-chain install risk (no external downloads or package installs). The skill uses only Python stdlib urllib/hmac/hashlib; nothing fetched at install time.
Credentials
The code requires AccessKeyId and SecretAccessKey for Volcengine stored in a local secrets.json, but the skill metadata declares no required env vars or config paths. Requesting cloud API credentials for this purpose is reasonable, but failing to declare them is a transparency/permission problem. The credentials are sensitive (can be used to call Volcengine) and the skill will read them from disk automatically when invoked.
Persistence & Privilege
The skill is not always:true and does not attempt to modify other skills or global settings. It writes output files to the agent workspace and skill-relative paths (temp_output.png, output.png) which is expected behaviour for a generator; no persistent installation or background services are created.
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 xingxiaolu-jimeng - 安装完成后，直接呼叫该 Skill 的名称或使用
/xingxiaolu-jimeng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
v1.0.2
- 新增内置50+即梦提示词模板，覆盖电商、社交、商业设计、视频等场景
- 文档补充了 "提示词模板库" 及万能公式与示例，提高提示词撰写效率
- 提供 `references/prompts-templates.md` 作为模板库详细说明
- 补充进阶技巧指导，包括角色定位、锁定元素、风格先行等
- 其他文档细节优化与调整
v1.0.0
- Initial release of jimeng-ai skill for image and video generation using Volcengine Jimeng API.
- Supports text-to-image and image-to-video generation via Python scripts.
- Configurable output: save to file or display media URL for inline use in OpenClaw.
- Integrates with token-manager for secure key management.
- Includes workflow and parameter guidance for both command-line and OpenClaw integration.
- API summary and usage notes provided for quick reference.
元数据
常见问题
Jimeng AI Image Generator 是什么？
即梦AI生图与视频生成工具。通过火山引擎API调用即梦AI，支持文生图、图片生成、视频生成。内置50+提示词模板覆盖电商、社交、商业设计、视频等场景。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 65 次。
如何安装 Jimeng AI Image Generator？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install xingxiaolu-jimeng」即可一键安装，无需额外配置。
Jimeng AI Image Generator 是免费的吗？
是的，Jimeng AI Image Generator 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Jimeng AI Image Generator 支持哪些平台？
Jimeng AI Image Generator 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 Jimeng AI Image Generator？
由 luis1213899（@luis1213899）开发并维护，当前版本 v1.0.2。
推荐 Skills