← 返回 Skills 市场
97
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install genealogy-agent
功能描述
Extract, structure, research, and visualize family history from raw text. Builds knowledge graphs, generates Mermaid trees, Obsidian vaults, and GEDCOM exports.
使用说明 (SKILL.md)
\r \r
genealogy-agent\r
\r A powerful agent skill for extracting, structuring, researching, and visualizing family history from raw text (memoirs, archives, chat logs).\r \r
Features\r
- LLM Extraction: Parses unstructured text to find people, dates, locations, and relationships.\r
- Knowledge Graph: Stores family data in a local JSONL ontology graph.\r
- Obsidian Vault Generation: Automatically generates an Obsidian-compatible vault (Markdown files with YAML frontmatter) for your family tree.\r
- Auto-Research: Autonomously searches global and regional databases (FamilySearch, Ancestry, OBD Memorial, Geneteka, etc.) to find new facts about ancestors.\r
- GEDCOM Export: Exports your family tree to the standard
.gedformat for use in MyHeritage, FamilySearch, or any desktop genealogy software.\r - Mermaid Visualization: Automatically generates beautiful family tree diagrams.\r \r
Setup\r
- Install dependencies:
uv pip install pydantic litellm duckduckgo-search\r - Set your preferred LLM provider (e.g.,
OPENAI_API_KEYorGEMINI_API_KEY).\r \r
Usage\r
You can ask the agent to:\r
- "Extract family members from this text: [text]"\r
- "Build a family tree from the extracted data"\r
- "Research my great-grandfather Ivan Ivanov born in 1910"\r
- "Export my family tree to GEDCOM"\r
- "Generate an Obsidian vault from my family data"\r \r
Tools\r
extract_family_data: Extracts structured family data (people, dates, relationships) from raw text using an LLM.\rbuild_family_graph: Saves extracted data into the local ontology graph.\rgenerate_mermaid_tree: Generates a Mermaid diagram from the graph.\rgenerate_obsidian_vault: Generates an Obsidian-formatted vault from extracted family data.\rresearch_ancestor: Autonomously researches an ancestor using web search and LLM extraction.\rexport_gedcom: Exports extracted family data to a standard GEDCOM (.ged) file.
安全使用建议
This skill appears to implement the genealogy features it claims, but take these precautions before installing: 1) Check the 'uv' usage — the SKILL.md and skill.json use `uv run` and `uv pip install`, which is unusual; confirm what 'uv' refers to or correct it to the intended commands (likely plain `python`/`pip` or a documented runner). 2) Expect the research tool to perform web searches and to send extracted context to whichever LLM provider you configure — do not supply API keys if you are unwilling to send private family data to third-party services. 3) The manifest does not declare required environment variables (OPENAI_API_KEY/GEMINI_API_KEY) — ask the author to list them explicitly in the skill metadata. 4) If you care about privacy, run the scripts locally in an isolated environment (air-gapped or without LLM keys) or remove/disable the research component before use. 5) Review the scripts yourself (they are small and included); they perform only local file writes and web searches via duckduckgo_search, with no obvious obfuscated code or external upload endpoints. If the author can clarify the 'uv' commands and update the metadata to declare LLM credentials, the skill would be more coherent and easier to trust.
功能分析
Type: OpenClaw Skill
Name: genealogy-agent
Version: 1.0.1
The genealogy-agent bundle contains significant security vulnerabilities, primarily related to shell injection and path traversal. In `skill.json`, tool commands use unvalidated template parameters (e.g., `{{text}}`, `{{input_file}}`) directly in shell executions via `uv run`, which could allow an attacker to execute arbitrary commands. Furthermore, scripts such as `build_graph.py`, `generate_obsidian.py`, and `export_gedcom.py` write files to user-specified paths without sanitization, posing a path traversal risk. While the network access in `research.py` (DuckDuckGo) and the file operations align with the stated genealogy purpose, the lack of input validation makes the bundle risky for deployment.
能力评估
Purpose & Capability
Name/description align with the included scripts: extraction (extract.py), graph building, Mermaid/Obsidian generation, GEDCOM export, and autonomous research (research.py). The requested capabilities (LLM extraction and web search) are reasonable for genealogy. Minor mismatch: the manifest (skill.json) does not declare the LLM-related environment variables that the SKILL.md and scripts expect.
Instruction Scope
SKILL.md and scripts explicitly instruct the agent to call an LLM (litellm) and to perform web searches (duckduckgo_search) for 'auto-research'. That means user-provided PII (names, dates, places) will be sent to external LLM providers and queried against public websites. This behavior is consistent with the skill's stated 'Auto-Research' feature but is privacy-sensitive and should be highlighted to users.
Install Mechanism
This is instruction-only (no formal install spec) but SKILL.md tells users to install dependencies with the command `uv pip install pydantic litellm duckduckgo-search` and the manifest's tool commands use `uv run ...`. The 'uv' prefix is non-standard/unclear (likely a typo or dependency on an undocumented CLI). There is no declared install script or required binary named 'uv', which is an incoherence that could break execution or lead users to run unfamiliar commands.
Credentials
The code uses litellm and the SKILL.md tells users to set provider keys (OPENAI_API_KEY or GEMINI_API_KEY), but the skill metadata lists no required environment variables or primary credential. Asking for LLM API keys is proportionate to the stated functionality, but the manifest omission is misleading and increases the chance a user will inadvertently leak family PII to third-party LLMs without realizing the skill requires those keys.
Persistence & Privilege
The skill does not request persistent/always-on privileges (always:false), does not modify other skills, and has no declared system-wide config changes. It operates via invoked scripts and local file I/O only.
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 genealogy-agent - 安装完成后,直接呼叫该 Skill 的名称或使用
/genealogy-agent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
v1.0.1
Clean republish: unified single commit, no leftover references.
v1.0.0
Initial ClawHub publication: extract family data from text, build knowledge graphs, auto-research ancestors, generate Mermaid trees, Obsidian vaults, and GEDCOM exports.
元数据
常见问题
Genealogy Agent 是什么?
Extract, structure, research, and visualize family history from raw text. Builds knowledge graphs, generates Mermaid trees, Obsidian vaults, and GEDCOM exports. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 97 次。
如何安装 Genealogy Agent?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install genealogy-agent」即可一键安装,无需额外配置。
Genealogy Agent 是免费的吗?
是的,Genealogy Agent 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
Genealogy Agent 支持哪些平台?
Genealogy Agent 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 Genealogy Agent?
由 Flo(@flobo3)开发并维护,当前版本 v1.0.1。
推荐 Skills