← 返回 Skills 市场
mar-docstrange
作者
MarjorieBroad
· GitHub ↗
· v1.0.0
· MIT-0
67
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install mar-docstrange
功能描述
Document extraction via SkillBoss API Hub. Convert PDFs and images to markdown, JSON, or CSV with confidence scoring. Use when you need to OCR documents, ext...
安全使用建议
This skill appears to do what it says (it posts documents to an external SkillBoss API and expects an API key). Before installing: 1) Verify the skill's origin and the correct homepage/contact for SkillBoss (registry metadata lacked this). 2) Confirm the platform will require/provide SKILLBOSS_API_KEY as package.json and SKILL.md indicate. 3) Treat any documents you send as potentially sensitive — avoid uploading PII or secrets unless you trust SkillBoss's privacy/security posture and have read their policy. 4) Prefer providing the API key via the agent's secret store or environment variable, not a plain config file; if you must store it in ~/.openclaw/openclaw.json follow the provided file-permission advice and rotate keys regularly. 5) Test the skill with non-sensitive documents first. If you cannot verify the vendor or the metadata mismatch, consider not installing or contact the platform support for provenance confirmation.
能力标签
能力评估
Purpose & Capability
SKILL.md and package.json consistently describe a document-extraction integration with https://api.heybossai.com and require SKILLBOSS_API_KEY — this aligns with the stated purpose. However, the registry metadata shown above lists no required env vars and no homepage; that discrepancy between registry metadata and the included files is unexplained and worth verifying (source is 'unknown').
Instruction Scope
Runtime instructions are limited to calling the external API (POST to https://api.heybossai.com/v1/run) and uploading documents (via URL or base64). The instructions do not ask the agent to read unrelated local files, other credentials, or system state. They do explicitly instruct storing and using an API key and optionally storing it in OpenClaw config.
Install Mechanism
This is an instruction-only skill with no install spec and no code files to execute locally, which minimizes install-time risk. package.json is present but there is no installer or archive download step.
Credentials
Only one credential is required in the SKILL.md and package.json: SKILLBOSS_API_KEY (declared as primaryEnv). That is proportionate for a cloud OCR API. The concern is the registry metadata at the top reported 'Required env vars: none' — inconsistent with the files. Confirm which declaration the platform will enforce before providing secrets.
Persistence & Privilege
The skill does not request always:true, does not modify other skills or system-wide settings, and is user-invocable. It does not request elevated or persistent system privileges.
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 mar-docstrange - 安装完成后，直接呼叫该 Skill 的名称或使用
/mar-docstrange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
v1.0.0
Initial release of DocStrange skill for document extraction via SkillBoss API Hub.
- Supports conversion of PDFs and images to markdown, JSON, or CSV with per-field confidence scores.
- Allows extraction of invoice fields, receipt parsing, and table conversion to structured data.
- Features both synchronous (≤5 pages) and asynchronous (>5 pages) extraction workflows.
- Offers advanced options: bounding boxes, hierarchy output, financial documents mode, and custom extraction instructions.
- Includes detailed setup, security, and best practice documentation.
元数据
常见问题
mar-docstrange 是什么？
Document extraction via SkillBoss API Hub. Convert PDFs and images to markdown, JSON, or CSV with confidence scoring. Use when you need to OCR documents, ext... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 67 次。
如何安装 mar-docstrange？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install mar-docstrange」即可一键安装，无需额外配置。
mar-docstrange 是免费的吗？
是的，mar-docstrange 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
mar-docstrange 支持哪些平台？
mar-docstrange 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 mar-docstrange？
由 MarjorieBroad（@marjoriebroad）开发并维护，当前版本 v1.0.0。
推荐 Skills