Hypothesis Generation
spacering-net/codeg
Structured hypothesis formulation from observations. An agent skill from spacering-net/codeg.
Analyzes four pillars (Bazi) charts with a research-style procedure: layered rules, graded evidence, stated uncertainty and predictions that can be tested.
SKILL.md written in Chinese; this summary is our English description.
$ npx skills add qianye-wuyu/yueyuan-bazi --skill yueyuan -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install qianye-wuyu/yueyuan-bazi yueyuan --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
Claude Code skills documentation · loads skills from .claude/skills/
Install the "yueyuan" agent skill from https://github.com/qianye-wuyu/yueyuan-bazi/tree/main into .claude/skills/yueyuan/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "yueyuan", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add qianye-wuyu/yueyuan-bazi --skill yueyuan -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install qianye-wuyu/yueyuan-bazi yueyuan --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "yueyuan" agent skill from https://github.com/qianye-wuyu/yueyuan-bazi/tree/main into .agents/skills/yueyuan/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "yueyuan", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add qianye-wuyu/yueyuan-bazi --skill yueyuan -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install qianye-wuyu/yueyuan-bazi yueyuan --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "yueyuan" agent skill from https://github.com/qianye-wuyu/yueyuan-bazi/tree/main into .cursor/skills/yueyuan/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "yueyuan", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add qianye-wuyu/yueyuan-bazi --skill yueyuan -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install qianye-wuyu/yueyuan-bazi yueyuan --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "yueyuan" agent skill from https://github.com/qianye-wuyu/yueyuan-bazi/tree/main into .gemini/skills/yueyuan/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "yueyuan", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install qianye-wuyu/yueyuan-bazi yueyuanInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add qianye-wuyu/yueyuan-bazi --skill yueyuan -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "yueyuan" agent skill from https://github.com/qianye-wuyu/yueyuan-bazi/tree/main into .github/skills/yueyuan/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "yueyuan", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add qianye-wuyu/yueyuan-bazi --skill yueyuan -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install qianye-wuyu/yueyuan-bazi yueyuan --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "yueyuan" agent skill from https://github.com/qianye-wuyu/yueyuan-bazi/tree/main into .opencode/skills/yueyuan/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "yueyuan", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
yueyuanAnalyzes four pillars (Bazi) charts with a research-style procedure: layered rules, graded evidence, stated uncertainty and predictions that can be tested.
This skill is a professional Bazi (four pillars) chart analysis system for Chinese metaphysics, built on the classical Ziping texts with a cross-check from the blind school. It is framed as a research procedure rather than a quiz solver: it allows uncertainty, refuses to flatter the user and expects claims that could be proven wrong. It does not draw images and has no teaching module. The skill text is in Chinese.
Analysis follows a hypothesis, evidence and counter-evidence loop, with layered rules and a fixed order for settling conflicts between them, evidence grades from A to E, competing hypotheses with stated uncertainty, a six-dimension consistency check on the favorable element and multi-school review. Predictions can be registered in advance and validated in two blind phases. Topics include chart verification, strength and weakness, structure patterns, ten gods, six relatives, luck cycles, yearly timing, and marriage, career, wealth and health.
When bazi_engine.py, which relies on lunar-python, is available, birth data is charted through its paipan function. Otherwise the agent falls back on a chart produced by a tool you supply. A references folder holds patterns, structure-change rules, six-relative notes and an HTML output template.
5 steps, taken from the first numbered list in SKILL.md.
Read from SKILL.md and the folder at commit 723b73c. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Ships script files (Python, from the files we listed), which the agent can run.
From the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Yueyuan Bazi Analysis loads about 8k tokens when it runs, and up to ~35k if it reads all its reference files. Until then it costs about 82 tokens; SKILL.md has 1,352 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
The full file from qianye-wuyu/yueyuan-bazi at commit 723b73c, republished under its MIT licence (© qianye-wuyu). 1,352 words, ~8,009 tokens.
.claude/skills/yueyuan/SKILL.md (or your agent's skills folder). This skill also uses 136 other files; get the full folder from GitHub.本 skill 是纯分析型专业八字系统:研究程序而非判题程序,允许不确定、拒绝迎合、要求可证伪。不做画图取象,不含教学模块。 v1.1 修订:规则分层+冲突裁决、删除伪量化(月令 50%/岁运加倍/神煞放大)、用神六维(新增专旺/从化)、地支改"初始检视顺序"、墓库 P0 修正、预注册预测+后验四态、合议两轴。 v1.2 修订(Gemini + Qwen 双模型审阅,独立裁决):通根双维分离(气量深度 vs 贴身感应)、变格/外格术语统一+顺势翻转机制、跨体系冲突裁决(子平主框架 vs 盲派)、竞争假设锚点绑定、Phase A/B 防污染隔离协议、输出模式控制(标准/深度)、从格阈值去量化、调候"力减"改状态判定、墓库"开启/坏根"判定闭环、调候兜底纲领。 v1.3 修订(GPT 最终评审,独立裁决):变格外格取消统一翻转、拆为 NORMAL/FROM/ZHUANWANG/HUAGQI 独立状态机(新库
references/biange.md)、极端气候引入 CLIMATE OVERRIDE 特殊裁决状态(解决调候与 CORE/SCHOOL 层级冲突)、调候优先级写死(专项 > 月令分区 > 通用兜底)、证据分级新增 S 层(输入可信度)、Phase A/B 文档化为两次独立调用契约(如实标注能力边界)、时柱未知修正(夫妻宫仍可论)、独立信号定义、收敛措辞分级(「大概率」→「目前更支持」)、建禄羊刃与变格体系分组、四吉四凶限定为格局功能语言、HTML 输出契约、人元司令改为可选算法、从格「无本气根」改证据之一(Root resistance)、病型组合论库(patterns.md)。 v1.4 修订(独立审批清单 + 独立裁决):竞争假设触发门槛(路线级竞争仅限四类临界盘)、不确定度口径(高/中/低档位判据)、从格逐条判据清单、references 静态库「待核对」标注 + 核对总则、Phase A/B 局限前置 + 强防迎合引导句、HTML 固定骨架模板(html-template.md)、通根「伪量化 vs 定性层级」区分、examples 脱敏样例。 v1.5 修订(第二轮审批清单 + 独立裁决,命理内容查漏补缺):相害=相穿合并(同一关系非两种)、枭神夺食概念修正(特指偏印夺食神,正印克伤官≠枭神夺食)、印重身弱/埋金拆分为「印多为病」+「母多灭子」、食神制杀/杀印相生/羊刃驾杀移出病型表改「成格范式」、新增病型(七杀无制/杀旺印弱/刃旺无制/贪合忘生/贪合忘克)、六亲补兄弟=比劫、比劫夺财补比肩劫财轻重、财多身弱补逢运反转、官杀混杂补食神制杀留官+去留判据、调候表覆盖度声明、专旺纯度要求、空亡查法引导。 v1.6 修订(做题复盘方法论沉淀,独立审批):七批复盘规则(R1–R59,58 条)落地——大运-流年分层与优劣判断、信号链/烈度三档/候选年排序、婚姻三层分离与婚离五档信号链、六亲取象口径修正、财富任财校验、健康取象双条件与程度四档、学业/性格板块、红线 4 项(第 5/7 条扩展 + 职业禁则 + 体貌禁则)。 v1.7 修订(应试策略剥离、回归研究程序,独立审批):从主流程剥离「排除法 / 取平凡 / 候选年排序选答案 / 两类错误分流 / 三态平凡先验 / 学制现实校验」等为四选一答题设计的规则,改写为「分析→倾向+证据+不确定度+验证方法」的研究程序语言;新增「分析纪律」小节(防跑题 / 防误用 / 按所问取象强制);命理规律表述面向用户(六亲取象、应期信号链、婚姻三层分离、健康取象双条件等命理内核全部保留);反推时辰降级为「生时不确定用户」的专用功能、不进主流程;测验专用内容归入 tests/,不参与用户分析。 v1.8 修订共两批(同版本号,批一=命理修正、批二=引擎接入): 【批一】开放分析命理修正——从格制化闸门 / 印星量级 / 财官反例 / 应期翻转 / 格局层次与职业分化,独立审批):①从格「制化自救」闸门——极弱日主先查可用制化(食神制杀/印星化杀/伤官合杀/合绊护身),制化神透干得地则优先正格制化不判从格,从格须「制化机制被破或无可用制化」为前提(对齐袁树珊食神制杀、梁湘润休废口径);②印星量级→身旺/身弱闸门——印星成势须过日主根气+财官克耗两道,裁决印为用/印为病;③喜用「财官」反例优先序——身弱财官不一律列忌,排查财旺生官/财印两用/印重喜财损印三路径,火食伤制杀不降辅助;④应期「吉凶翻转」对称判据——财官旺运须区分忌神攻身 vs 用神得地化杀/损印,吉则成就窗口、忌则压力关口;⑤格局层次+职业取象分化——成格断高位/显贵、破格断中下/平实,层次与旺衰强弱不直接挂钩;身强杀旺区分得用掌权/假神得用偏业,从杀格须再核制化。校准样本(抽象化):例12/13(从格)、例5/9(印星)、例4/6/7/9/10(应期)、例13/15(职业取象)。详见references/biange.md、references/patterns.md。 【批二】引擎接入,2026-08-17 独立审批:排盘约定改为「内置引擎优先」——运行环境提供bazi_engine.py(lunar-python)时,出生信息直接调用paipan()排盘;无引擎环境再退回用户提供排盘工具结果。仅排盘流程层修改,不动任何命理规则。 方法论来源:三路并行检索 + 多源交叉验证(《渊海子平》《三命通会》《滴天髓》《子平真诠》《穷通宝鉴》《神峰通考》 + 现代格局派/旺衰调候派/盲派),流派分歧处显式标注,不混用。
两条独立通道:S 层 = 输入可信度(八字本身对不对),A–E 层 = 推理证据(基于该输入的判断依据)。分开标注,如「S1/A」。
| 等级 | 内容 | 说明 |
|---|---|---|
| S3 | 输入可信度最高 | 出生资料 + 时区 + 节气 + 地点均确认,或两个独立排盘工具结果一致 |
| S2 | 输入可信度较高 | 排盘软件导出,单一工具 |
| S1 | 输入可信度一般 | 用户直接手打四柱,未注明来源 |
| S0 | 输入可信度存疑 | 手打 + 干支不合法/自相矛盾,须先核对再分析 |
| A | 原局硬数据 | 四柱干支、藏干、月令、日主(以引擎输出或用户提供的已排盘为准——引擎可用时按第四章优先级取引擎输出,可信度另以 S 层标注) |
| B | 结构事实 | 合、冲、刑、害(穿)等关系是否存在(存在≠生效) |
| C | 命理判断 | 「三合是否化金」「某神是否得根」等需论证的判断,必须给依据 |
| D | 流年应验 | 用户提供的人生事件,只能作后验检验证据 |
| E | 模型解释 | 因果解读,属假设性说明,与事实性内容分开呈现 |
铁则:C 不得伪装成 B;E 不得伪装成 A/D。S0 输入不得直接进入 A 级分析,必须先请用户核对排盘。
| 标签 | 层级 | 示例 | 硬度 |
|---|---|---|---|
[STRUCTURAL] | 结构事实 | 四柱、藏干、合冲刑害、通根透藏 | 最高,不可改 |
[CORE] | 核心理论 | 月令定权、格局、扶抑、病药(经典共识层) | 高 |
[SCHOOL] | 流派理论 | 人元司令分野、调候(穷通宝鉴系)、盲派做功、阳干根深阴干根浅 | 中,须标注流派 |
[HEURISTIC] | 经验规则 | 岁运并临强化、墓库冲开、应期常见范式 | 初始参考,不作硬断 |
[SUPPLEMENTAL] | 辅助符号 | 神煞、空亡 | 只作主题标签/修正项 |
[EVENT] | 事件解释 | 婚恋/事业/财富/身心倾向 | 输出层,须带不确定度 |
[PROHIBITED] | 红线 | 见第七章 | 违反即输出无效 |
冲突裁决规则:
STRUCTURAL > CORE > SCHOOL > HEURISTIC > SUPPLEMENTAL,高一层胜出(此为裁决优先级,非五行力量量化)本模型不做手工排盘(不在语言模型里心算干支/起运——干支排算与起运计算交给确定性计算层,出错即全盘皆错)。
bazi_engine.py(依赖 lunar-python,见 requirements.txt),调用 paipan(year, month, day, hour, gender, minute=0) 生成四柱 / 大运 / 起运 / 藏干 / 十神 / 空亡 / 十二长生 / 天月二德等结构化输入,分析基于引擎输出。mingli_rules.py 同层)。「引擎可用」判据 = 当前环境能成功 import bazi_engine;尝试 import 失败(无 lunar-python / 无引擎文件 / 无法执行 Python)即视为引擎不可用,自动降级。bazi_engine.py::paipan() 排盘,以引擎输出为准| # | 校验项 | 通过条件 | 不通过处置 |
|---|---|---|---|
| 1 | 四柱干支合法 | 每柱 2 字、干支有效、阳干配阳支/阴干配阴支 | 提示核对,不强行分析 |
| 2 | 时辰已知 | 给到时柱 | 降级:年月日三柱 + 标注「时柱未知——日柱夫妻宫仍可分析;涉及子女、晚年及部分时柱应期的结论降级/待补时柱后深化」 |
| 3 | 性别已知 | 男/女 | 降级:中性分析 + 大运不排,标注「补性别后补运」 |
| 4 | 大运已知 | 起运岁/运干支 | 无则标注「大运缺失」:引擎可用时补排,引擎不可用时需排盘工具提供 |
列出:四柱、日主(日干+五行阴阳)、月令提纲、地支藏干(本气/中气/余气)、空亡(以引擎/排盘工具所列为准)、日主十二长生。本步骤不做吉凶判断。
paipan() 已输出空亡(年/日柱)与日主十二长生;引擎不可用或用户提供盘面时,以排盘工具/用户提供为准;均缺失时显式标注「已略」,不臆补,见 references/suijie-shensha.md)。references/biange.md 对应状态机取用决策链(非单线透干即取格):
月令 → 藏干结构 → 透出情况 → 是否多神同透?
├─ 是 → 按本气优先、次看中气余气(多神同透取透出有力者,与 `mingli_rules.py::ding_ge()` 取格顺序一致);分辨正格与兼格,定主格
└─ 否 → 专取本气(不透取余气,均不透取本气)
→ 判定格局大类(A 正格 / B 特殊格局 / C 变格体系)→ 成格还是破格? → 有无救应? → 最终主格references/biange.md 判定,禁止共用「统一翻转器」(v1.3 P0)references/biange.md 第二节 FROM 前置条件「判据 0 · 制化自救闸门」。校准样本(抽象化):例12(含合绊护身之制化,梁湘润判休废不从)、例13(透食神制杀有根,袁树珊判食神制杀正格)。references/biange.md 中该状态自己的顺势/逆势/破格规则取用(FROM 喜克泄、ZHUANWANG 喜印比、HUAGQI 喜化神),不执行统一「喜忌整体翻转」;变格被破时明确宣告退论普通格局并说明喜忌变化mingli_rules.py 的 TIAO_HOU 代码表(120 格全表)为准,速查表为缩略参考,冲突时以代码表为准并提示版本差异,与规则约束.md 同口径)CORE > SCHOOL 生效)分别评估并给出一级、二级用神:
references/biange.md 该状态自己的顺势/逆势/破格规则取用:FROM 从势→喜克泄耗、忌生扶;ZHUANWANG 专旺→喜印比、忌财官食伤;HUAGQI 化气→喜化神五行、忌破化五行。先于普通扶抑判定;禁止跨状态套用喜忌)references/patterns.md 第八节):paipan() 已输出各支神煞——桃花/驿马/华盖/将星、日干神煞——天乙贵人/文昌/羊刃,及按月支的天月二德;引擎不可用或用户提供盘面时,以排盘工具/速查表为准)。统一边界(v1.6,R33,待验证):神煞只作主题标签/细节提示,禁止断人生状态、宗教归属、婚姻状态、桃花旺否、伴侣特征(桃花、华盖、寡宿同此约束)。作用:领域规则匹配前先锁定「用户问的是什么」,避免跑题、误用、取象冲突。这是强制步骤,不是可选项。
references/quxiang-mode.md);② 执行两项执行层防错检查(根气一致性 / 透干-暗藏术语一致性,见 references/analysis-checklist.md);③ 输出头部声明「取象模式:…;流程检查:根气一致性 PASS / 术语一致性 PASS」。本登记与检查属执行层防错(怎么描述、用哪条口径),不改变任何命理规则判断方向。铁则:「原局定事、岁运定时」——原局有象,岁运引动方应验。单一信号不足为凭,须 ≥2 个独立信号互证(如配偶星动 + 夫妻宫动 + 岁运合冲)。 「独立」定义(v1.3):独立 = 信号来自不同结构层——十神星动(如官星透)、宫位动(如夫妻宫被冲合)、岁运引动(大运流年与原局的合冲)、原局结构响应(如伏吟/反吟/天克地冲),四个层中取 ≥2 个不同层互证;同一层内用不同语言描述同一变量不算独立(如「配偶星透 + 财星透」常是同一件事) 前置/完成二分与独立互证同时生效(v1.6):完成信号年须以 ≥2 个不同结构层互证为准(完成信号自身通常已含岁运引动层,是否还需叠加宫动层按各事项细规则执行)。
| 档位 | 判据 | 事件落点 |
|---|---|---|
| 状态年/程度级 | 未伤星(仅宫动/墓逢值/伏吟/会局攻身) | 压力、反复、小变动 |
| 中事件 | 伤星(星透被克损/被合走/根被拔) | 破财、受伤、离异、失业、流产 |
| 大事件 | 星伤 + 库开 + 宫动三链齐备 | 亡故、牢狱、重疾 |
references/liuqin.md(六亲专库)。references/patterns.md 第九节)——修正「成格大贵统一断中产/层级中」的量级保守偏差。校准样本(抽象化):例1/2/4/5/6/9 成格大贵,断高位倾向而非一律中产;[HEURISTIC] 待验证。references/patterns.md 第九节 9.2)——【防污染隔离协议】(最高优先级,杜绝后见之明):
<Phase_A_Blind_Analysis> 与 <Phase_B_Posterior_Validation>),Phase A 区块完成前严禁引用、暗示任何经历数据;并如实告知用户:「单次会话内模型已看到经历,隔离为输出层约束,效果弱于真正分步执行」——这是诚实边界,不伪装成数据隔离。Phase A · 预注册(在任何用户经历之前完成,输出后锁定):
最优假设:…
主要用神:…
关键争议:…
未来 2–3 个高概率主题:…
每个主题的证据:…
什么情况会推翻:…Phase B · 后验验证(用户提供经历后执行):
【阶段性收敛结论】(核心特色,用户第一次给反馈后强制输出):
设计动机:多数用户不愿持续提供完整验证数据,常只给「某一年某一事」就停止。若此时仍停在「部分命中、待验证」,会让用户觉得「又要我查资料」而失去耐心。正确做法是先给阶段性甜头,再温和说明还需什么。
【喜忌翻译与领域重解读】(强制):
设计动机:喜用神对非专业用户是抽象符号——「喜火木水」四个字无法直接与生活挂钩。只要喜忌方向明确,就必须立刻用这把「钥匙」打开门后的内容,让用户拿到的不是「一个喜用神」,而是「基于我喜用神的完整人生解读」。
references/wuxing-life.md 给出 1–2 条最突出的生活落地建议(方位/颜色/环境/职业意象),措辞温和、标注民俗文化层,不做恐吓式「忌神绝对避免」建议。每个重大判断(旺衰、格局、用神、应期)统一使用 Hypothesis 模板,且先列"路线级"假设,再列"参数级"假设。
触发门槛(v1.4,防止「永远不确定」):
触发锚点(强制内嵌,不得文末补票):该模板必须内嵌于对应步骤的输出中——步骤 1(旺衰)、步骤 2(格局)、步骤 5(用神)、步骤 8(应期)出现路线级分歧时,即时在那一节输出模板,不得脱离步骤、在全文末尾统一堆砌。
假设 A:…
支持证据(标注等级):…
反证:…
预测差异:若未来 X 岁运出现 Y 则更支持 A;出现 Z 则支持 B
假设 B:…(成立条件 / 当前缺什么 / 反证)
当前最优 + 不确定度:…bazi_engine.py)或用户提供/排盘工具的结果,发现疑点提示用户核对references/aesthetics.md + 套用 references/html-template.md 固定骨架:Design Tokens、五行五色、●喜○忌、预期管理提示、结论先行、区块固定顺序),HTML 是本系统的标志性交付物。执行契约(如实):环境支持文件生成 → 真正产出 HTML 文件(套用固定模板,非逐盘手写);环境不支持 → 输出 Markdown 等价结构并说明「本环境未生成 HTML 文件,以下为等价 Markdown」;不得声称「已生成 HTML」而实际只有文字。仅当用户明确只要文字答案、或为单点问答(非完整分析)时,才省略 HTML。{{坤造/乾造}} 处)必须与用户输入性别一致——男 → 乾造、女 → 坤造,二选一、禁止并列、禁止反向。性别以用户明确提供的输入为准(引擎排盘时以 paipan() 输出的 性别 字段为准);性别缺失时先向用户确认,不得自行推断。产出后自检:若头部性别与正文取象口径(男=财为妻/官杀为子女,女=官杀为夫/食伤为子女)不一致,视为输出无效、须重写。references/wuxing-life.md 给出 1–2 条最突出的生活落地建议(方位/颜色/环境/职业意象),措辞温和、标注民俗文化层、不做恐吓式「忌神绝对避免」建议。SKILL.md 为主执行协议;references/ 为规则资料库。缺失 reference 时不得臆造,对应模块降级或标注「资料不可用」。 阶段说明:阶段一 = 排盘/计算引擎(
bazi_engine.py,确定性计算);阶段二 = 规则库(mingli_rules.py调候 120 格 + 旺衰四指标 + 月令定格、规则约束.md分析约束、案例库/真实命例参照),为分析提供确定性输入与真实命例参照。 测验专用内容(v1.7):大赛题复盘 / 做题规则 / 两类错误分流等为四选一答题设计的工具见tests/,不参与用户分析;本 skill 只对真实用户的八字做研究型分析。 静态知识库核对总则(v1.4):references 中所有静态表格/歌诀/古籍章句(人元司令表、调候表、神煞歌诀等),凡未在文件内标注「已核对」的,一律视为「待核对」;使用时若命中相关条目,应提示用户「该条目可能存在版本差异,已按待核对处理」,不得作为绝对依据。禁止凭模型记忆臆造或改写古籍内容。
references/biange.md:变格体系规则库——NORMAL/FROM/ZHUANWANG/HUAGQI 四态独立状态机([CORE]/[SCHOOL],v1.3 新增;v1.8 补 FROM「判据 0 · 制化自救闸门」:极弱日主先查制化/合绊护身,制化神透干得地则优先正格制化、从格须制化被破)references/patterns.md:病型与组合论速查——经典病型(枭神夺食/伤官见官/财坏印/比劫夺财/财多身弱/印多为病/母多灭子/官杀混杂/七杀无制/杀旺印弱/刃旺无制/贪合忘生/贪合忘克)+ 经典成格范式(伤官佩印/食神制杀/杀印相生/羊刃驾杀)+ 印星喜忌先决校验([CORE]/[HEURISTIC],v1.5 重构)+ 第八节财官喜忌反例裁决 + 第九节格局层次与职业取象分化(v1.8 新增)references/liuqin.md:六亲专库——六亲取象统一口径(R1–R9)、配偶星条款(R26/R27)、双星互动断法(R56)、宫位分工表(年柱=祖辈、月柱=父母、日支=夫妻、时柱=子女)、病型×六亲断语分级表(v1.6 已落地,[CORE]/[HEURISTIC])references/quxiang-mode.md:取象模式契约——分析流程第一步登记 CORE-DEFAULT / SCHOOL-SOURCE / DUAL 三态(仅用户显式指定原批/流派才可进入 SCHOOL-SOURCE,不覆盖全局 CORE;由「分析纪律」第 4 条认领执行)references/analysis-checklist.md:执行层防错清单——根气一致性 + 透干/暗藏术语一致性两项检查(每盘必查;只防执行层错误、不改命理规则;由「分析纪律」第 4 条认领执行)references/r-index.md:R 编号总索引(R1–R59 主题与出处速查;R13/R35 未使用)references/renyuan-command.md:人元司令分野表 + 旺衰条件变量细则([SCHOOL],可选算法)references/tiaohou-notes.md:《穷通宝鉴》十干十二月核心调候速查 + 决策树([SCHOOL])references/suijie-shensha.md:岁运应期法则 + 空亡用法 + 神煞七种速查([HEURISTIC]/[SUPPLEMENTAL])references/wuxing-life.md:五行喜用的生活应用指南——方位/颜色/环境/职业意象([HEURISTIC]/[SUPPLEMENTAL],文化传统层,不参与判断核心)references/aesthetics.md:输出审美规范(内置 Taste Enhancer 原则,[SUPPLEMENTAL])——制作 HTML/图表/卡片时强制遵循references/html-template.md:HTML 报告固定骨架模板(Design Tokens + 区块结构,[SUPPLEMENTAL],v1.4 新增)——产出 HTML 时强制套用../mingli_rules.py:阶段二规则库(代码规则,位于工作空间根目录)——调候用神 120 条决策表(tiao_hou_yongshen())+ 旺衰四指标(analyze_wangshuai())+ 月令定格(ding_ge(),含建禄/羊刃)。为步骤 1/2/4 提供确定性输入;调候表状态为「待核对」,命中时提示版本差异。../规则约束.md:阶段二规则约束文本(位于工作空间根目录)——旺衰综合权重 + 八格成败救应 + 用神取用三模块 + 输出纪律。LLM 分析时须遵守,与本文档步骤 1/2/4/5 对齐。../案例库/案例库.md:阶段二案例库(位于工作空间根目录 案例库/ 子目录)——15 例真实经典命例(古籍 + 现代名家),每例含命盘/生平/格局/用神病药/大运应验/六亲/出处/流派标注。作分析参照示例,不作训练数据;与 tests/cases/ 虚构测试盘区分,虚构盘不得作为分析依据引用。其余案例库文件(v3 封存集、女命候选集、新15例、补充10例)为评估/验证专用,不参与用户分析参照。© qianye-wuyu, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
SKILL.md and 136 other files (references) in the repository root of qianye-wuyu/yueyuan-bazi.
Open the folder on GitHubat commit 723b73c
Yueyuan Bazi Analysis next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Yueyuan Bazi Analysis this skillqianye-wuyu/yueyuan-bazi | 266 | — | ~8k | Automated safety check: Pass | MIT | |
| Hypothesis Generationspacering-net/codeg | 3.9k | 14 repos | ~3.6k | Automated safety check: Notes | MIT | |
| GitHub Deep Researchbytedance/deer-flow | 84k | 4 repos | ~1.3k | Automated safety check: Pass | MIT | |
| Nature Paper CardYuan1z0825/nature-skills | 47k | 2 repos | ~2.1k | Automated safety check: Pass | Apache-2.0 | |
| DfAM Printability Checkearthtojake/text-to-cad | 19k | 2 repos | ~1.5k | Automated safety check: Pass | MIT | |
| Content Research Writerweapp-tailwindcss/weapp-tailwindcss | 1.9k | 25 repos | ~3.5k | Automated safety check: Pass | MIT |
spacering-net/codeg
Structured hypothesis formulation from observations. An agent skill from spacering-net/codeg.
bytedance/deer-flow
Researches a GitHub repository over four rounds using the GitHub API and web search, then writes a structured markdown report with timeline, metrics and Mermaid diagrams.
Yuan1z0825/nature-skills
Builds a structured deep-reading card for one scientific paper, covering methods, how experiments support claims, limitations and research ideas, with a script to prepare the source.
earthtojake/text-to-cad
Measures STL, OBJ, PLY or 3MF mesh files against additive manufacturing design limits and reports printability findings for each print process.
weapp-tailwindcss/weapp-tailwindcss
Assists in writing high-quality content by conducting research, adding citations, improving hooks, iterating on outlines, and providing real-time feedback on each section.
mvanhorn/last30days-skill
Research what people actually say about any topic in the last 30 days.
Categories
Analyzes four pillars (Bazi) charts with a research-style procedure: layered rules, graded evidence, stated uncertainty and predictions that can be tested. This skill is a professional Bazi (four pillars) chart analysis system for Chinese metaphysics, built on the classical Ziping texts with a cross-check from the blind school. It is framed as a research procedure rather than a quiz solver: it allows uncertainty, refuses to flatter the user and expects claims that could be proven wrong.
Yueyuan Bazi Analysis fits situations like: analyzing a Bazi chart from a birth date and time; judging strength, structure and favorable elements in a four pillars chart; reviewing luck cycles and yearly timing for a chart; comparing two charts for marriage compatibility.
Run `npx skills add qianye-wuyu/yueyuan-bazi --skill yueyuan -a claude-code`. Or copy the skill folder (the qianye-wuyu/yueyuan-bazi repository) into .claude/skills/yueyuan in your project. Claude Code loads it when a task matches its description.
Run `npx skills add qianye-wuyu/yueyuan-bazi --skill yueyuan -a codex`. Or copy the skill folder (the qianye-wuyu/yueyuan-bazi repository) into .agents/skills/yueyuan in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add qianye-wuyu/yueyuan-bazi --skill yueyuan -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/yueyuan, .gemini/skills/yueyuan, .github/skills/yueyuan and .opencode/skills/yueyuan in your project.
Going by SKILL.md and its folder, Yueyuan Bazi Analysis needs Python for the scripts in its folder. Our summary lists: Birth date and time, or a chart from your own charting tool.
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
Yueyuan Bazi Analysis is published under the MIT licence (from the LICENSE file in the skill folder). It allows redistribution, so the full SKILL.md is shown on this page.
About 8k tokens (SKILL.md is roughly 32k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 27k tokens, read only when the agent opens those files.
Skills that share tags, products or a category with Yueyuan Bazi Analysis: Hypothesis Generation (spacering-net/codeg, 3.9k stars), GitHub Deep Research (bytedance/deer-flow, 84k stars), Nature Paper Card (Yuan1z0825/nature-skills, 47k stars) and DfAM Printability Check (earthtojake/text-to-cad, 19k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
qianye-wuyu (a GitHub user) maintains it in qianye-wuyu/yueyuan-bazi, which has 266 GitHub stars. The repository was last updated on August 22, 2026.
Source: qianye-wuyu/yueyuan-bazi on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.