中文说明在前，English follows.
vedic-astro-pyjhora-skill 是一个面向 Codex / agent 工作流的吠陀占星技能包。它基于 CNWU16/vedic-astro-skills 的 v5 五模块架构，并结合 PyJHora 改装为自动排盘版本。
这个版本的目标是：用户把本仓库下载下来交给 Codex 或 Claude Code，让 agent 自动安装技能、补齐依赖并执行完整流程。用户只需要提供出生日期、时间和地点，agent 会通过 PyJHora 生成 structured_data.md，再进入 reader 校验、core P1-P12 审计、career/love 专项分析与 rectifier 校时流程。
- PyJHora 自动排盘：无需手工导出 Jagannatha Hora PDF。
- 五模块架构：
reader / core / career / love / rectifier。 - 结构化输出：生成可被后续分析模块复用的
structured_data.md。 - 16 条校验：SAV/BAV、Rahu-Ketu、D9 映射、Dasha 顺序等。
- 自动依赖安装：脚本发现缺少依赖时会尝试自动安装。
- 时间校准：内置
time_scan.py，支持上升与 D9 边界扫描。 - 报告打包：将 Markdown 分析文件自动打包为 HTML。
本项目主要不是给用户手动运行 Python 脚本，而是给 Claude Code / Codex / OpenCode 这类 agent 使用。
最简单的方式是：把这个仓库链接直接发给 agent：
https://github.com/singkitlee/vedic-astro-pyjhora-skill
然后让它自动下载、安装、补依赖、运行。
请打开并安装这个开源仓库里的 vedic-astro skill：
https://github.com/singkitlee/vedic-astro-pyjhora-skill
请你自动完成以下事情：
1. 下载或读取该仓库；
2. 把 skills/vedic-astro 安装到当前环境可识别的 skills 目录；
3. 自动安装运行所需的 Python 依赖；
4. 使用 PyJHora 自动排盘，不需要我手动导出 JHora PDF；
5. 如果出生时间靠近上升或 D9 边界，请先进行校时；
6. 最后按 reader -> core -> career/love/rectifier 的流程生成完整 Markdown/HTML 报告。
我的出生信息是：
姓名：
性别：
出生日期：
出生时间：
出生地点：
当前最想问的问题：
本地测试中，Claude Code + Opus 4.7 的组合使用效果不错，尤其适合执行长流程：读盘、校验、校时、分段写报告、再打包 HTML。Codex 和 OpenCode 也可以按同样方式使用。
推荐步骤：
- 下载或克隆本仓库。
- 在 Codex 或 Claude Code 中打开这个仓库目录。
- 直接对 agent 说：
请安装这个仓库里的 vedic-astro skill，并用它帮我排盘/读盘。
- agent 应自动把
skills/vedic-astro安装到对应的 skills 目录，并在第一次运行时自动安装 Python 依赖。
如果你希望自己安装到 Codex 全局 skills 目录，可以执行：
git clone https://github.com/singkitlee/vedic-astro-pyjhora-skill.git
cd vedic-astro-pyjhora-skill
mkdir -p ~/.codex/skills
cp -a skills/vedic-astro ~/.codex/skills/vedic-astroClaude Code 用户可让 Claude Code 根据本仓库的 skills/vedic-astro/SKILL.md 自动完成安装；不同安装位置会因本地配置而不同。
请使用 vedic-astro skill。
出生信息：1990年5月15日14:30，出生地 Ningbo, Zhejiang, China，性别男。
请先自动排盘和校验，如果时间靠近上升/D9边界，请先校时，再继续完整解盘。
请把这个仓库里的 skills/vedic-astro 安装到 Codex 或 Claude Code 可识别的 skills 目录。
安装后用 PyJHora 自动排盘，不需要我手动导出 JHora PDF。
python skills/vedic-astro/scripts/chart_generator.py \
--date "1990-05-15" \
--time "14:30" \
--place "Ningbo, Zhejiang, China" \
--output ./demo/parts/structured_data.md \
--style structured \
--time-accuracy "精确到分钟"python skills/vedic-astro/scripts/time_scan.py \
--date 1990-05-15 \
--time 06:30 \
--lat 29.8699867 \
--lon 121.5427313 \
--range 30 \
--save ./demo/parts/rectification_scan.mdpython skills/vedic-astro/scripts/report_builder.py ./demo --lang cn一般用户不需要直接运行这些脚本；它们主要供 agent 自动调用，或供开发者调试。
本项目明确关联并感谢两个开源项目：
- CNWU16/vedic-astro-skills：提供 v5 五模块架构、reader/core/career/love/rectifier 工作流思想、报告组织方式与校验体系基础。
- naturalstupid/pyjhora：提供吠陀占星计算能力，本项目的自动排盘脚本以 PyJHora 为核心计算引擎。
本项目不是上述项目的官方版本，也不代表上游维护者背书。
特别感谢 CNWU16/vedic-astro-skills 项目。它把吠陀占星分析拆成 reader、core、career、love、rectifier 五个清晰模块，并把数据校验、验前事、P1-P12 审计、专项分析和报告打包组织成可执行的 agent 工作流。本项目的结构化思路、模块边界和许多分析流程都受其启发。
在此基础上，本项目主要做了 PyJHora 自动排盘适配，让用户可以从出生信息直接生成结构化盘面，减少手工导出和 OCR 误差。感谢 CNWU16 对这个方向的开源探索，让后续改装有了清楚、扎实的起点。
本项目使用 GNU Affero General Public License v3.0 or later (AGPL-3.0-or-later)。
选择 AGPL 的原因是：本项目与 PyJHora 组合使用，而 PyJHora 使用 AGPL；同时 AGPL 能确保代码和网络服务修改保持开放，符合本项目“尽可能开放”的发布目标。
vedic-astro-pyjhora-skill is a Vedic astrology skill package for Codex / agent workflows. It is based on the v5 five-module architecture of CNWU16/vedic-astro-skills, and adapts it with PyJHora to support automatic chart generation from birth data.
The goal of this project is simple: users download this repository and hand it to Codex or Claude Code. The agent installs the skill, installs missing dependencies, generates structured_data.md through PyJHora, then continues into reader validation, core P1-P12 audit, career/love deep dives, and rectifier-based birth-time calibration.
- PyJHora-powered chart generation: no manual Jagannatha Hora PDF export required.
- Five-module workflow:
reader / core / career / love / rectifier. - Structured output: reusable
structured_data.mdfor downstream modules. - Validation checks: SAV/BAV, Rahu-Ketu axis, D9 mapping, Dasha sequence, and more.
- Automatic dependency installation when script dependencies are missing.
- Birth-time rectification support through
time_scan.py. - Markdown-to-HTML report packaging.
This project is primarily designed for Claude Code / Codex / OpenCode agents, not for end users to manually run Python scripts.
The simplest way is to send this repository URL directly to your agent:
https://github.com/singkitlee/vedic-astro-pyjhora-skill
Then ask it to download, install, install dependencies, and run the workflow automatically.
Please open and install the vedic-astro skill from this open-source repository:
https://github.com/singkitlee/vedic-astro-pyjhora-skill
Please automatically:
1. Download or read the repository;
2. Install skills/vedic-astro into the skills directory recognized by this environment;
3. Install all required Python dependencies;
4. Use PyJHora for automatic chart generation, without asking me to export a JHora PDF manually;
5. If the birth time is near a Lagna or D9 boundary, run birth-time rectification first;
6. Then follow the reader -> core -> career/love/rectifier workflow and generate complete Markdown/HTML reports.
My birth data:
Name:
Gender:
Birth date:
Birth time:
Birth place:
Main question:
In local testing, Claude Code + Opus 4.7 has worked well for this workflow, especially for long-running tasks such as chart extraction, validation, rectification, staged report writing, and HTML packaging. Codex and OpenCode can use the same workflow.
Recommended flow:
- Download or clone this repository.
- Open the repository folder in Codex or Claude Code.
- Ask the agent:
Please install the vedic-astro skill from this repository and use it for my Vedic chart reading.
- The agent should install
skills/vedic-astrointo the appropriate skills directory and automatically install Python dependencies on first run.
For Codex global skills:
git clone https://github.com/singkitlee/vedic-astro-pyjhora-skill.git
cd vedic-astro-pyjhora-skill
mkdir -p ~/.codex/skills
cp -a skills/vedic-astro ~/.codex/skills/vedic-astroClaude Code users can ask Claude Code to install the skill from skills/vedic-astro/SKILL.md; the exact destination may vary by local setup.
Please use the vedic-astro skill.
Birth data: May 15, 1990, 14:30, Ningbo, Zhejiang, China, male.
Please generate the chart, validate it, run rectification if the time is near a Lagna/D9 boundary, then continue into the full reading.
Please install skills/vedic-astro from this repository into the skills directory recognized by Codex or Claude Code.
After installation, use PyJHora for automatic chart generation. I do not want to manually export a JHora PDF.
python skills/vedic-astro/scripts/chart_generator.py \
--date "1990-05-15" \
--time "14:30" \
--place "Ningbo, Zhejiang, China" \
--output ./demo/parts/structured_data.md \
--style structured \
--time-accuracy "exact to the minute"python skills/vedic-astro/scripts/report_builder.py ./demo --lang enMost users do not need to run these commands manually. They are mainly intended for agent automation and developer debugging.
This project explicitly acknowledges and builds upon two open-source projects:
- CNWU16/vedic-astro-skills: the source of the v5 five-module workflow, validation-first design, P1-P12 audit structure, specialized career/love modules, and report organization ideas.
- naturalstupid/pyjhora: the Vedic astrology calculation engine used by this project's automatic chart generation scripts.
This project is not an official release of either upstream project and is not endorsed by their maintainers.
Special thanks to CNWU16/vedic-astro-skills. Its v5 architecture turns Vedic astrology analysis into a clear agent workflow: reader, core, career, love, and rectifier. Its work on validation, pre-reading verification, P1-P12 audit, specialized analysis, and report packaging provides the conceptual foundation for this project.
This repository extends that foundation by integrating PyJHora-based automatic chart generation, so users can start from birth data directly and reduce manual export or OCR errors. CNWU16's open-source work made this adaptation possible.
This project is licensed under the GNU Affero General Public License v3.0 or later (AGPL-3.0-or-later).
AGPL was chosen because this project integrates with PyJHora, which is AGPL-licensed, and because AGPL keeps both source-code modifications and network-service use open to the community.