Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Horosa 星阙 — Astrology & Chinese Metaphysics

    1

    Give your AI 110 Horosa techniques: local engines, confirmed settings, sourced knowledge, traceable exports and structured reports. 让 AI 真正调用星阙:110 技法、本机计算、参数确认、引据解读与结构化报告。

    Free

    1 useSecurity scanned
    Horosa 星阙 — Astrology & Chinese Metaphysics

    Horosa 星阙 — Astrology & Chinese Metaphysics

    Example session with this skill installed

    {
      "natal": {
        "date": "1990-07-15",
        "time": "23:10:00",
        "zone": "-03:00",
        "lat": "34s36",
        "lon": "58w22",
        "hsys": 3,
        "zodiacal": 1,
        "siderealAyanamsa": "raman",
        "ad": 1
      },
      "returnsOpts": {
        "startYear": 2024,
        "count": 5
      }
    }
    
    • Read your context and instructions
    • Compiled the horosa

    Historical public regression fixture: the engine response was captured on 24 September 2026, then rendered by the upstream snapshot builder with the export clock fixed at 2026-09-24 02:01. The complete stored snapshot below is reproduced verbatim; this is not a new calculation or a personal consultation.
    公开历史回归夹具:引擎响应于 2026-09-24 实抓,再由上游快照构建器生成;导出时钟固定为 2026-09-24 02:01。下方保留完整快照原文,不是本次新计算或个人咨询结论。

    [起盘信息]
    出生时间:1990-07-15 23:10:00 周日
    经纬度:58w22 34s36
    时区:-03:00
    黄道:恒星黄道·Raman
    宫制:Placidus
    盘型:夜生盘

    [太阳/月亮返照时间轴]
    区间:2024 年起 5 年(太阳返照 = 太阳回到本命度;首个月亮返照 = 该年首个月亮回本命度)

    年份太阳返照首个月亮返照太阳返照上升月亮返照上升
    20242024-07-17 04:39:242024-01-18 11:20:53金牛 15.24°双鱼 10.90°
    20252025-07-17 10:52:232025-01-08 01:09:25室女 0.05°天秤 10.21°
    20262026-07-17 16:58:452026-01-25 21:09:26射手 19.97°狮子 6.03°
    20272027-07-17 23:09:332027-01-16 08:02:31双鱼 6.94°摩羯 28.05°
    20282028-07-17 05:26:132028-01-06 11:36:38金牛 25.17°双鱼 4.26°

    [当前时点]
    导出时刻:2026-09-24 02:01
    盘主当前年龄:36.19 岁

    [方法说明]
    回归轴:逐年列太阳返照(太阳回本命度)与该年首个月亮返照时刻及两盘上升点。
    读法:返照上升落座定该年/该月主色,上升与本命宫位的对应指示焦点领域;多年并列可见上升轮转的节律。

    Source / 公开来源: https://github.com/Horace-Maxwell/horosa-skill/blob/v0.41.1/horosa-skill/tests/fixtures/sync311_newtools_live.json

    Connects securely to your tools. The creator never sees your data.

    About this skill

    110 techniques. A complete workflow from calculation to report.

    Bring the depth of Horosa 星阙 into your AI agent: Western natal and predictive astrology, returns, primary directions, horary and electional work—alongside BaZi, Zi Wei Dou Shu, Da Liu Ren, Qimen, Taiyi, Jinkou, Liu Yao, tarot and the 14 Shen Shu systems.

    The engine calculates. The agent explains from the actual results. Horosa connects specialist computation to a disciplined reading workflow:

    • Confirm what matters. Resolve result-changing settings before calling a technique.
    • Keep the evidence. Read structured export snapshots and disclose missing sections, method choices, computation sources and version provenance.
    • Go deeper with sources. Consult the cited methodology knowledge base; distinguish sourced doctrine from general interpretation.
    • Compare perspectives. Build multi-technique synthesis from evidence-linked templates, with disagreements and setting conflicts visible.
    • Keep the work. Search local run history, revisit earlier results and turn readings into DOCX, PDF or JSON reports.

    After installation, the default calculation workflow runs locally. This skill supplies the agent guide and reference sheets; it requires the Horosa MCP server or CLI and offline runtime. Optional cloud integrations change the data flow when enabled. Most chart exports use Chinese terminology.

    110 个技法,把星阙的专业计算能力交给你的 AI。

    从西洋本命盘、返照、推运、主限、卜卦与择日,到八字、紫微、大六壬、奇门、太乙、金口诀、六爻、塔罗及十四路神数——把星阙的专业计算能力接入你的 AI 工作流。

    先确认参数,再调用真实引擎;每一步解读,都能回到实际盘面。

    • 参数有交代: 时间、地点、历法和流派口径等关键设置,确认后再算。
    • 结果有依据: 读取结构化导出,附技法口径、算源与版本信息;缺段和降级如实披露。
    • 方法有出处: 调用带引用的知识库,分清教义依据与一般推断。
    • 合参见分歧: 多路技法共同分析,结论绑定各自证据,保留不一致与口径冲突。
    • 成果能积累: 本地检索历史盘面与解读,导出 Word、PDF 或 JSON,形成可回看的研究与咨询资料。

    适合重视方法深度、参数透明和结果留存的占星师、术数研究者与进阶使用者。 默认计算在本机完成;安装本 skill 后,还需配置 Horosa 本地服务与离线 runtime。

    The Markdown skill package does not contain the calculation runtime. Install and connect Horosa before requesting calculations.
    No native offline payload is provided for Linux or Intel Macs; they require a configured gateway to a supported host.
    The Windows runtime directory must use an ASCII-only path. Reserve approximately 5 GB for installation and working space.
    Missing result-changing settings require clarification. The agent must not invent them or bypass confirmation gates.
    Some output sections depend on the tool, chosen settings or runtime availability. Non-empty warnings indicate incomplete results and must be disclosed.
    Multi-technique tools provide comparison templates; the agent supplies the interpretation. Chart computation does not establish the accuracy of divinatory predictions.
    Most export text and specialist terminology are Chinese. Fully English output is not guaranteed.
    There is no hosted Horosa endpoint. Web-based assistants require a separately configured authenticated HTTPS gateway.
    The project documents GitHub wheel installation; its PyPI distribution channel is not yet enabled.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 26 days ago

    • Free with an Agensi account

    Requires

    Horosa Skill 0.41.1
    uv (Python tool runner)
    Python 3.12+
    Horosa offline runtime

    Permissions

    Terminal / Shell
    Network Access
    Write Files
    Read Files
    Environment Variables

    File Scopes

    ~/.horosa-skill/**
    ~/.horosa/runtime/**
    %APPDATA%/HorosaSkill/**
    %LOCALAPPDATA%/Horosa/runtime/**

    The agent invokes local Horosa tools or CLI commands. The runtime reads installed engines and local records, writes SQLite run history, JSON artifacts, diagnostic traces and requested DOCX/PDF/JSON reports. Initial setup and upgrades download software from GitHub and package sources. Onboarding can also update the selected client's Horosa MCP entry and skill folder, with backups; it preserves unrelated client settings. Local backend defaults are 127.0.0.1:9999 and 127.0.0.1:8899; the optional HTTP MCP listener defaults to 127.0.0.1:8765. Default calculations are local. TypeSafe Jev is off by default; enabling it can send redacted question metadata, or export-snapshot text under an explicitly selected snapshot scope. HOROSA_TRACE_OTLP_ENDPOINT is unset by default; setting it exports diagnostic events to the configured endpoint. Trace capture of payloads and AI answers is off by default. Remote gateways, mirrors and other optional destinations use the user's configuration. The AI client's own data handling also applies. No browser, email, contact, microphone or camera access is required for the core workflow.

    Listed27 days ago

    What's inside

    Frequently Asked Questions

    No. It guides an agent to use the Horosa MCP server or CLI. Install the Horosa package and offline runtime, then connect them to your agent.

    Local runtime packages are available for Apple Silicon Macs and Windows x64. Windows 11 ARM runs the x64 package through emulation. Linux and Intel Macs use gateway mode with a supported host.

    Yes, after the runtime is installed, with the default local configuration. Installation, updates and explicitly enabled remote integrations require network access.

    Time, location, time zone, house system, calendar rules and method settings can change results. The skill confirms missing result-sensitive settings instead of inventing them.

    Runs, questions, AI answers and reports are stored locally under the configured Horosa data directory. The default is ~/.horosa-skill on macOS and %APPDATA%/HorosaSkill on Windows. Optional cloud decisions, remote gateways and diagnostic export endpoints change the data flow only when configured.

    Yes. Horosa supports local history search and DOCX, PDF and JSON reports. Reports use actual tool results and the agent's supplied interpretation.

    No. The project supplies a local MCP/CLI runtime. Browser-based assistants require a separately configured authenticated gateway.

    The skill is released under AGPL-3.0-only. The included LICENSE and public source repository contain the full terms. The separate Horosa runtime bundles third-party ken engines that retain their own MIT license notices.

    New in Research & Analysis