← Back to Skills Marketplace
760
Downloads
3
Stars
2
Active Installs
2
Versions
Install in OpenClaw
/install jimeng-skills
Description
基于火山引擎即梦 AI 生成图片和视频，并将结果整理为适合 OpenClaw 在飞书、企微、钉钉中发送和查看的纯文本结果、远程直链和可公开分享链接。用于“生成图片”“生成视频”“让飞书/企微/钉钉可查看”“输出适合企业 IM 的分享结果”“调用 jimeng-skills”等场景。
README (SKILL.md)
jimeng-skills
使用 skill.yaml 里声明的脚本入口，不要直接假设富文本卡片能力可用。
工作原则
- 优先返回纯文本摘要，避免依赖平台专有卡片协议。
- 始终返回本地文件路径，便于调试和二次处理。
- 能拿到火山引擎返回的远程 URL 时，一并返回远程直链。
- 如果设置了
JIMENG_PUBLIC_BASE_URL，额外返回公开分享 URL。企业 IM 中优先发送这些https://...链接。
输出策略
- 图片任务：返回
remoteUrls、publicUrls、localFiles。 - 视频任务：返回
remoteUrls、publicUrls、localFiles。 - 任务未完成时：返回
status: pending和便于用户重试的说明。 - 文本结果放在
text字段中，内容使用简单换行和编号，不使用 Markdown 表格。
推荐调用
图片：
npx ts-node scripts/openclaw-jimeng.ts image "一只戴墨镜的柴犬，电影海报风格"
视频：
npx ts-node scripts/openclaw-jimeng.ts video "一只戴墨镜的柴犬在沙滩上奔跑" --wait
公开分享 URL
若需要飞书、企微、钉钉直接打开本地下载文件，先把输出目录映射到你自己的静态文件服务，再设置：
export JIMENG_PUBLIC_BASE_URL="https://your-domain.example.com/jimeng-output"
脚本会把本地 output/... 路径转换成对应的公开 URL。
注意
VOLCENGINE_AK必填。- 永久凭证使用
VOLCENGINE_SK；临时凭证使用VOLCENGINE_TOKEN。 - 企业 IM 对本地绝对路径不可见，跨设备查看时必须依赖远程直链或
JIMENG_PUBLIC_BASE_URL。
Usage Guidance
This skill's instructions expect Volcengine credentials and local TypeScript scripts, but the package metadata lists no required env vars and contains no code. Before installing or running it: (1) Ask the publisher for the missing script files (scripts/openclaw-jimeng.ts and any dependent code) or a verified source repository; do not run npx commands that you don't trust, because npx can download and execute code from npm. (2) Verify and only provide VOLCENGINE_AK and a short-lived VOLCENGINE_TOKEN where possible; avoid pasting permanent SK into an unverified skill. (3) If you need JIMENG_PUBLIC_BASE_URL, ensure the static host you point to is under your control and that publishing files there is acceptable for your data. (4) Request that the skill metadata be corrected to declare required env vars and include provenance (homepage, source repo, or owner contact). With those fixes and bundled code you can reassess; in the current state the mismatches are suspicious.
Capability Analysis
Type: OpenClaw Skill
Name: jimeng-skills
Version: 0.1.1
The provided files consist of metadata and documentation for a skill designed to generate images and videos via Volcengine's Jimeng AI. The SKILL.md file provides functional instructions for the AI agent on how to format outputs for enterprise IM platforms (Feishu, WeCom, DingTalk) and manage local/remote URLs. No malicious code, exfiltration patterns, or harmful prompt injection attempts were found in the provided content.
Capability Assessment
Purpose & Capability
The skill's stated purpose (generate images/videos via 火山引擎 / Volcengine) legitimately requires Volcengine credentials (AK/SK or temporary token) and possibly a public base URL for sharing; SKILL.md explicitly names VOLCENGINE_AK and VOLCENGINE_SK / VOLCENGINE_TOKEN. However, the registry metadata lists no required environment variables or primary credential — that mismatch is incoherent and could lead to hidden runtime prompts or failures.
Instruction Scope
The instructions reference running scripts (npx ts-node scripts/openclaw-jimeng.ts) and returning local file paths, remote URLs, and public share links. But this package contains no script files or code; it's instruction-only. That means the skill either expects external code not bundled here or assumes an environment with those scripts present, which is a significant operational mismatch and could cause the agent to attempt to download or run code from npm or other sources.
Install Mechanism
There is no install spec (instruction-only), so nothing is written to disk by an installer from this bundle. This reduces direct install risk, but it increases runtime ambiguity since the agent may run commands (e.g., npx) that fetch code at execution time.
Credentials
The environment variables named in SKILL.md (VOLCENGINE_AK, VOLCENGINE_SK or VOLCENGINE_TOKEN, and optional JIMENG_PUBLIC_BASE_URL) are proportionate to the claimed purpose. The concern is that these required env vars are not declared in the skill metadata — the agent or install process may request them interactively or assume they exist, which is ambiguous and risky. Also, the docs recommend permanent SK; advise using temporary tokens where possible.
Persistence & Privilege
The skill does not request elevated persistence (always: false) and does not declare system-wide config or other skills' config changes. Autonomous invocation is allowed (platform default) but is not combined with other high privileges here.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install jimeng-skills - After installation, invoke the skill by name or use
/jimeng-skills - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.1
Version 0.1.1 of jimeng-skills
- No file changes detected in this release.
- No updates to features, documentation, or behavior.
v0.1.0
Initial release of jimeng-skills:
- Generate images and videos using Volcano Engine's Jimeng AI and organize results for sharing in enterprise IM platforms (Feishu, WeCom, DingTalk).
- Outputs pure text summaries prioritizing remote URLs, public share URLs, and local file paths.
- Unified output strategy for image and video tasks, including status handling and easy-to-retry instructions.
- Clearly documented environment variable setup for public sharing.
- Platform-agnostic output, avoiding proprietary rich card protocols.
Metadata
Frequently Asked Questions
What is Jimeng Skills?
基于火山引擎即梦 AI 生成图片和视频，并将结果整理为适合 OpenClaw 在飞书、企微、钉钉中发送和查看的纯文本结果、远程直链和可公开分享链接。用于“生成图片”“生成视频”“让飞书/企微/钉钉可查看”“输出适合企业 IM 的分享结果”“调用 jimeng-skills”等场景。 It is an AI Agent Skill for Claude Code / OpenClaw, with 760 downloads so far.
How do I install Jimeng Skills?
Run "/install jimeng-skills" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Jimeng Skills free?
Yes, Jimeng Skills is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Jimeng Skills support?
Jimeng Skills is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Jimeng Skills?
It is built and maintained by klaus1534 (@klaus1534); the current version is v0.1.1.
More Skills