← 返回 Skills 市场
337
总下载
0
收藏
0
当前安装
7
版本数
在 OpenClaw 中安装
/install open-card
功能描述
从本地 OpenClaw数据生成或迭代 OpenCard 个人名片。读取工作区身份信息、配置和会话统计；渲染 HTML 预览；导出 PNG。只读模式——不写入用户档案数据。
安全使用建议
This skill appears to do what it says: read your local OpenClaw workspace, prepare material for an LLM to write two text fields, render an HTML preview, and (optionally) export a PNG via Playwright + a local Chromium. Before installing or running it: 1) Inspect your ~/.openclaw USER.md / IDENTITY.md / MEMORY.md for sensitive content you don't want included, because excerpts (up to the documented lengths) will be supplied to whatever model the agent uses — if your agent uses a cloud model those excerpts will be transmitted to that provider. 2) Be aware the default background image and Google Fonts are fetched from the network during rendering, which exposes your client IP; you can override the background with a local file to avoid that. 3) PNG export requires installing playwright-core and a local Chrome/Chromium; installing npm packages globally is optional but required for export. 4) If privacy is a concern, run generation with a local model or remove/clean sensitive files before use. The code is readable and contains explicit safeguards (no hidden endpoints, no message-body extraction), so the main risks are data exposure via the documented LLM/model path and browser resource requests.
功能分析
Type: OpenClaw Skill
Name: open-card
Version: 1.0.6
The open-card skill bundle is a legitimate tool designed to generate personal business cards from local OpenClaw data. While it accesses sensitive files such as session logs and identity documents (~/.openclaw/workspace/), it does so transparently to extract statistics and context for the card's content. The scripts demonstrate security awareness, using HTML escaping in render-background-card.py and avoiding string interpolation for shell commands in export-background-card.sh. Network activity is limited to fetching legitimate UI assets (Google Fonts and a background image from Cloudflare R2), and the documentation explicitly details the data flow and privacy considerations.
能力评估
Purpose & Capability
Name/description say: read local OpenClaw data, render HTML preview, export PNG. Declared binaries (python3, bash, node, npm), an npm dependency (playwright-core), and the included scripts (collect-data.py, render-background-card.py, export-background-card.sh) are consistent with that purpose.
Instruction Scope
SKILL.md and scripts explicitly read files under ~/.openclaw (workspace files, openclaw.json, sessions/*.jsonl). The code extracts excerpts (USER.md / IDENTITY.md / MEMORY.md up to specified lengths) and token/platform stats; it avoids reading message bodies and does not call external APIs itself. However the browser-based rendering will request external resources (a Cloudflare R2 background image URL and Google Fonts) which will reveal client IP and HTTP metadata; this is documented in the SKILL.md. Also, the excerpts intended for LLM generation will be sent to the model provider if the agent uses a cloud model — this data flow is documented but is an important privacy consideration.
Install Mechanism
Install spec only declares 'playwright-core' via npm (a legitimate package for headless Chromium automation). No arbitrary download URLs, no obscure installers, and the scripts do not auto-download executables; Chrome/Chromium must be present locally. Installing playwright-core is moderate-risk (npm package) but appropriate and proportional for PNG export.
Credentials
The skill requests no environment variables or credentials. It does read local files in the home directory which may contain sensitive text; excerpts of USER.md/IDENTITY.md/MEMORY.md are intentionally collected to feed the LLM. This is proportionate to the stated feature but users should be aware those excerpts will reach any cloud LLM the agent uses.
Persistence & Privilege
The skill is not always-enabled, does not request elevated system privileges, and does not modify other skills' configs. It writes only temporary files (temporary JSON/preview paths outside the skill directory per instructions) and does not persist new profile data by default.
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 open-card - 安装完成后，直接呼叫该 Skill 的名称或使用
/open-card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
v1.0.6
- 新增 metadata 字段，声明运行依赖和 npm 安装建议（如 playwright-core）。
- 明确依赖项 node、npm、playwright-core 的安装方式，更便于自动化部署。
- 运行依赖说明部分同步，将 playwright-core 依赖描述与 metadata 一致。
- 其他内容和使用说明无变动，保持核心流程与数据安全边界一致。
v1.0.5
- 新增 born_label 和 born_date 字段，右下角展示 OpenClaw 的部署时间，替代原有 timezone 字段。
- 平台字段 platforms 显式归一化：CLI/WEB/各平台名，按用户可理解方式映射，去除内部来源。
- 精简 LLM 生成字段，仅 recent_focus 和 openclaw_review，去除 timezone 生成相关说明。
- schema 结构调整，完善 born_label、born_date 字段说明与回填逻辑。
- 其他部分文档与字段说明同步更新，更贴近实际数据采集和展现效果。
v1.0.4
open-card 1.0.4
- 明确说明网络通信细节，新增“网络请求与数据流向”章节，区分浏览器资源访问与 LLM 上下文传递。
- 说明 LLM 文案生成过程涉及本地文件摘录，如使用云端模型，这些摘录会随上下文发往模型 API。
- 强调脚本自身不直接发出 API 请求，模型调用由 agent 运行时负责。
- 增补关于凭证管理与模型访问权限界定的说明，明确本 skill 不读取任何 API 密钥。
- 更新会话数据隐私条款，提醒用户根据隐私需求选择本地或云端模型。
- 其余功能及安全边界未变。
v1.0.3
**Changelog for open-card v1.0.3**
- Major documentation overhaul: skill instructions and field schema rewritten in Chinese for clarity and specificity; all workflow and rendering guidelines updated.
- Data extraction and field generation process clarified: AI agent now generates `timezone`, `recent_focus`, and `openclaw_review` from unified `copy_inputs` context, not from direct file reads.
- Schema and field naming updated; usage period standardized to `token_30d` for count fields, and card data sample updated accordingly.
- File structure simplified: added `references/field-schema.md`, and removed outdated reference docs.
- Privacy, data handling, and network request sections updated for improved transparency and user safety.
v1.0.2
- Adds requirements for three key LLM-generated fields: `recent_focus`, `openclaw_comment`, and `openclaw_review`, including detailed generation rules and writing guidance for each.
- Documents network requests made during HTML/image rendering (background image, Google Fonts), and clarifies that no user data is transmitted externally.
- Lists runtime dependencies for data collection, HTML preview, and PNG export, with fallback/error guidance.
- Updates the example output schema to include `openclaw_comment` and `openclaw_review`.
- Refines instructions for copywriting tone and content, prioritizing qualitative insights about the user based on local files.
- Other descriptive and process clarifications for improved safety, visual quality, and clarity in workflow and guardrails.
v1.0.1
**OpenCard 1.0.1 Changelog**
- Removed the premium card render workflow and template files (`render-card.py`, `premium-template.html`).
- Updated documentation to standardize on the background-image card rendering process only.
- Clarified use of `render-background-card.py` and `background-template.html` for all output.
- Export instructions now reference only the background-image composition, simplifying the workflow.
- Cleaned resources section to match available scripts and templates.
v1.0.0
- Initial release of OpenCard: generate a shareable OpenClaw profile/namecard using only existing local agent data.
- Reads workspace identity files and OpenClaw runtime to extract and normalize profile information.
- Renders a visual HTML preview before exporting as an image; supports collaborative iteration on design and layout.
- Strictly read-only: does not invent or persist new user fields unless explicitly requested by the user.
- Follows clear visual and privacy guardrails—omits unreliable fields, uses clean labels, and avoids sensitive/internal details.
元数据
常见问题
openclaw's digital card 是什么？
从本地 OpenClaw数据生成或迭代 OpenCard 个人名片。读取工作区身份信息、配置和会话统计；渲染 HTML 预览；导出 PNG。只读模式——不写入用户档案数据。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 337 次。
如何安装 openclaw's digital card？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install open-card」即可一键安装，无需额外配置。
openclaw's digital card 是免费的吗？
是的，openclaw's digital card 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
openclaw's digital card 支持哪些平台？
openclaw's digital card 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 openclaw's digital card？
由 TrenLiu（@timothyliew2）开发并维护，当前版本 v1.0.6。
推荐 Skills