← 返回 Skills 市场
140
总下载
0
收藏
1
当前安装
4
版本数
在 OpenClaw 中安装
/install weipaiban-skill
功能描述
微排版技能 - 通过 AI 自动创建微信图文作品，支持模板搜索、配色方案、文本生成和图片替换
安全使用建议
This skill appears internally consistent with its purpose. Before installing/using: (1) Only provide WEIPAIBAN_API_KEY to enable the core functionality; prefer creating a scoped API key if the service supports it. (2) If you want image generation, provide VOLCENGINE_* only when needed and prefer temporary STS tokens over long-lived keys. (3) The skill may download a ~4.7MB rembg model (~/.u2net/u2netp.onnx) if you enable background removal; rembg installation is only performed after you explicitly confirm the shown commands. (4) Generated images and any replacements are uploaded to the weipaiban CDN — review the privacy/terms of both weipaiban.cn and any image-generation provider (jimeng/火山引擎) before uploading sensitive content. (5) Small metadata inconsistency: WEIPAIBAN_API_BASE is referenced in the docs but not declared in the manifest optionalEnv — confirm with the publisher or treat WEIPAIBAN_API_BASE as an optional override if you need a custom API endpoint. (6) If you want extra safety, run the skill in a sandbox/container and clean the /tmp/weipaiban-task-* directories and ~/.u2net/ after use.
功能分析
Type: OpenClaw Skill
Name: weipaiban-skill
Version: 1.2.0
The weipaiban-skill bundle is a legitimate tool designed to automate the creation of WeChat articles using the Weipaiban platform and Jimeng AI. It demonstrates a high level of transparency and security awareness by implementing mandatory user confirmation points (⏸️) for sensitive actions such as API submissions, template selection, and the installation of optional dependencies like 'rembg'. The skill uses shell commands (curl, rembg) and handles API keys, but these are strictly scoped to the stated purpose and are well-documented in SKILL.md and references/runtime-dependencies.md. No evidence of malicious intent, data exfiltration, or unauthorized persistence was found.
能力标签
能力评估
Purpose & Capability
The skill is a 微排版 (WeChat-layout) assistant and requires a WEIPAIBAN_API_KEY to call the weipaiban.cn APIs (template search, clone, parser, asset upload). Optional env vars (VOLCENGINE_*) and the optional jimeng-ai/rembg tools are explicitly tied to the optional image-generation Step 8. The declared writes (/tmp/weipaiban-task-*/ and ~/.u2net/u2netp.onnx) are explained in the docs and are proportional to the task. One minor metadata mismatch: the SKILL.md treats WEIPAIBAN_API_BASE as an optional env var, but it is not listed in the top-level openclaw metadata optionalEnv — functionally this is harmless but is an inconsistency in the manifest.
Instruction Scope
The SKILL.md contains precise step-by-step instructions: it reads/writes only the task-dir JSON files and images, calls the weipaiban APIs with the declared API key, and uses optional jimeng/rembg flow only if the user opts in. It explicitly forbids auto-installation and mandates user confirmations at multiple interaction points. It does instruct the agent to download and analyze images (and to call external image-generation service via an optional skill) — this is required for the stated image-replacement feature and is described transparently.
Install Mechanism
There is no install spec (instruction-only skill), so nothing is automatically written to disk at install time. The docs describe optional user-driven installs (pip install rembg, installing the jimeng-ai skill via clawhub/find-skill) and clearly state the skill will not auto-run those installs without explicit user confirmation. This is proportionate and low-risk from an install perspective.
Credentials
Only WEIPAIBAN_API_KEY is required (primary credential) which matches the core service. VOLCENGINE_AK/SK/TOKEN are optional and documented as needed only for optional image-generation. No unrelated high-privilege credentials are requested. Note: WEIPAIBAN_API_BASE (custom API base) is used in the documentation but is not declared in the metadata optionalEnv list — a minor manifest inconsistency to be corrected. Overall the requested environment access is proportionate to the described capabilities.
Persistence & Privilege
always:false (not force-included). The skill writes only to its own temporary task directories and an optional model cache (~/.u2net/u2netp.onnx). It does not request to modify other skills or global agent settings. The documented persistent writes are limited in scope and explained.
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 weipaiban-skill - 安装完成后，直接呼叫该 Skill 的名称或使用
/weipaiban-skill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
v1.2.0
**v1.2.0 introduces enhanced execution safety and improved dependency transparency.**
- 增加“执行门禁（安全基线）”说明，明确强制确认及受控自动连续执行动作，提升交互安全性
- 新增“写入最小化说明”，详细说明本地缓存及中间文件写入范围
- 明确 Step 9 增加“最终提交前变更摘要确认”为强制交互点
- 细化外部依赖和缓存行为，完善任务目录的写入说明
- 文档内用词及结构小幅优化，提升可读性
v1.1.0
- 更新图片生成可选依赖：使用 VOLCENGINE_AK/SK/TOKEN 代替 JIMENG_ACCESS_KEY/SECRET_KEY
- SKILL.md、运行时依赖和图片生成步骤文档同步变更为 Volcengine 配置方式
- 可选依赖检测和降级逻辑调整为匹配 VOLCENGINE 环境变量要求
- 其他文档细节与提示内容适配上述凭据变动
v1.0.1
**Step 8 图片生成改为可选并增加运行环境自检**
- 新增详细依赖清单与降级说明，支持图片生成功能的缺省与跳过
- 交互点增加依赖检测，图片生成阶段（Step 8）支持缺失 jimeng-ai/凭据/rembg 时由用户选择跳过
- 优化 metadata，声明可选环境变量、外部依赖和技能写入目录
- 迁移依赖、运行环境和执行流程相关说明至 references/runtime-dependencies.md
- 删除了旧的 CHANGELOG.md 与 README.md，精简文档结构
v1.0.0
首次发布
元数据
常见问题
微排版 是什么？
微排版技能 - 通过 AI 自动创建微信图文作品，支持模板搜索、配色方案、文本生成和图片替换. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 140 次。
如何安装 微排版？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install weipaiban-skill」即可一键安装，无需额外配置。
微排版 是免费的吗？
是的，微排版 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
微排版 支持哪些平台？
微排版 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 微排版？
由 weipaiban（@weipaiban）开发并维护，当前版本 v1.2.0。
推荐 Skills