chanlun-trading-system
把缠论分析变成先定级别、再认结构、先写失效点的可执行研究流程。Use for A-share, HK-stock, ETF, index, or futures analysis with 中枢、分型、笔、线段、背驰、区间套、三类买卖点 and multi-level recursion, or for converting Chan theory into reproducible backtest rules. Research and study only; never issue stock tips or promise returns.
Install / Use
npx skills add noahnan-max/chanlun-trading-systemInstalls into whichever agent you are using.
SKILL.md
Installable skill definition
Quality Score
Category
Education & ResearchSupported Platforms
Tags
Our assessment of chanlun-trading-system
chanlun-trading-system scores 75/100 on our quality scale, 408th of 441 Education & Research skills we index.
Its SKILL.md is 5.9 KB long, well organised into 12 sections with 1 code example: a thorough specification that gives an agent plenty to work with.
It has no GitHub stars yet, so there is no community track record; judge it on its content.
Maintenance, license and trust
- The repository was last updated about 2 months ago, so chanlun-trading-system is actively maintained.
- Our last check on 2026-09-02 found the source still online.
- No license is declared. By default that means all rights are reserved: you can read it, but reusing or redistributing it is not clearly permitted. Ask the author before building on it commercially.
- Its trust signals score 80/100, with 2 cautions from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
chanlun-trading-system compared with similar skills
All 4 of these similar skills score higher than chanlun-trading-system; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| chanlun-trading-system (this skill)by noahnan-max | 75 | 0 | 2mo ago | SKILL.md |
| last30days-skillby mvanhorn | 100 | 64.2k | 1d ago | CLAUDE.md |
| algorithmic-artby anthropics | 100 | 177.9k | 22d ago | SKILL.md |
| pptxby anthropics | 100 | 177.9k | 22d ago | SKILL.md |
| designby nextlevelbuilder | 100 | 133.6k | 11d ago | SKILL.md |
Frequently asked questions
- How do I install chanlun-trading-system?
- Run
npx skills add noahnan-max/chanlun-trading-system. The install tabs above show the steps for each supported agent. - Which AI agents does chanlun-trading-system work with?
- It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
- Is chanlun-trading-system safe to use?
- It declares no license and scores 80/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is chanlun-trading-system still maintained?
- The repository was last updated about 2 months ago, so chanlun-trading-system is actively maintained.
Skill content
View source on GitHubname: chanlun-trading-system description: "把缠论分析变成先定级别、再认结构、先写失效点的可执行研究流程。Use for A-share, HK-stock, ETF, index, or futures analysis with 中枢、分型、笔、线段、背驰、区间套、三类买卖点 and multi-level recursion, or for converting Chan theory into reproducible backtest rules. Research and study only; never issue stock tips or promise returns."
Chanlun Trading System(缠论分析 Skill)
把缠中说禅(缠论)的技术分析体系,变成一套可执行、可复核、不自欺的看盘与研究流程:图形诊断 → 结构判定 → 信号识别 → 失效点 → 回测规则化。
⚠️ 免责声明:本 Skill 仅用于技术研究与学习。它不构成投资建议,不给具体买卖指令,不承诺任何收益。缠论是一套技术分析框架,存在主观性与失效风险,任何据此产生的盈亏由使用者自负。本 Skill 的最大价值在于强制纪律——让分析"先定级别、先认结构、先写失效点",而不是给你一个"必涨信号"。
它解决什么问题
缠论难学、易被滥用:很多人把"指标金叉/MACD 底背离"直接喊成"买点",把"突破"喊成"三买",结果在震荡里反复挨打(whipsaw)。本 Skill 用一套**门控(gate)**逼着分析走完整结构链,任何一步缺失就自动降级为"观察"而不是"确认买入"。
三种运行模式(必须先声明用的是哪种)
strict_chanlun:有足够的多级别 K 线/图形细节,能完整判断包含、分型、笔、线段、同级别中枢、背驰、低级别触发——不跳过递归结构链时才用。structure_proxy:能复现 K 线包含、分型、笔、类中枢重叠,但线段/走势类型做了简化。proxy_research:只有日线、波段或指标数据时用。必须明确标注"近似损失",绝不把代理信号当成严格的缠论原文定义。
默认姿态:从 strict_chanlun_audit(严格原文自检)起步,只有在点名缺了哪一步结构之后才降级。
不可妥协的规则(Non-Negotiable Rules)
- 永远先定操作级别。没有级别,就没有有效的买卖点。
- 区分
original_definition / segment_proxy / swing_proxy / kline_proxy / indicator_proxy五种定义口径。 - 结构优先:走势类型 → 级别 → 中枢 → 背驰/力度 → 买卖点 → 过滤器。
- 不把小级别背驰直接当成大级别反转,除非检验了"小转大"的条件。
- 每个入场都要写明:依据、触发、失效点、下一个观察点。
- MACD/RSI/成交量/换手/资金流/趋势线/均线/筹码等技术过滤器只调整信心、等待纪律或仓位,不定义缠论买卖点。
- 数据不足以做严格的分型/笔/线段/中枢时,明确降级为代理分析。
- 结构不清时,宁可等待,不强行交易。
- 有图(截图/书图/手绘)时,先读图,再依赖周边文字。
- 任何信号,先记录"什么情况下它失效",再讨论潜在收益。
- 没点名"前一个同级别趋势"或"下跌+中枢+下跌"的上下文、没识别被比较的同向走势之前,不喊一买/一卖。
- 没点名"一买候选 + 它的第一次回抽测试"之前,不喊二买/二卖。
- 没点名"中枢边界、离开、回拉、不回/再入"全部状态之前,不喊三买/三卖。
- 回测产信号时,必须暴露这些列:
trade_level / confirm_level / trigger_level / definition_mode / structure_completeness / approximation_loss。
严格原文自检门(Strict Original Audit Gate)
任何操作性结论之前,过这道门:
level_gate:点名review_level / trade_level / confirm_level / trigger_level。structure_gate:按"包含 → 分型 → 笔 → 线段 → 中枢"的顺序构建或引用。type_gate:把当前同级别状态归类为 趋势 / 盘整 / 中枢延伸 / 中枢突破 / 转折 / 不明。comparison_gate:用背驰时,点名被比较的两段同向走势,并确认它们同级别。buy_sell_gate:上述门通过后,才把结构映射到一/二/三类买卖点。trigger_gate:要求低级别触发,或显式标注trigger_missing。risk_gate:先写失效点,再谈收益。downgrade_gate:任一门不过,输出proxy_research或observe,而不是buy_confirmed。
被禁止的捷径:指标背离→一买;突破→无回抽的三买;更高的低点→无一买上下文的二买;30 分钟均线/MACD 状态→当成 30 分钟结构;日线信号+30 分钟指标确认→冒充严格多级别缠论信号。
默认工作流(8 步)
- 背景:确认标的、周期、目的、约束;设
trade/confirm/trigger级别(中线=周/日/30m;波段=日/30m/5m;短线=30m/5m/1m);声明输出是 strict / structure_proxy / proxy_research。 - 结构:有图有数就识别包含、分型、笔、线段、中枢、当前态;只有文字时给假设+清单,不假装看到图。(读图见
references/visual-reading.md,结构判定见references/structure-engine.md,概念冲突见references/concepts.md) - 状态:用一个当前态标签(
center_oscillation / center_extension / center_breakout / uptrend / downtrend / transition_zhongyin / first|second|third_buy_candidate / 卖侧镜像 / untradable_unclear),先点名最近有效中枢及其zg/zd边界。 - 信号:找候选买卖点并说明理由;背驰必须先点名被比较的两段。(细则见
references/buy-sell-playbooks.md,多级别/区间套/小转大见references/multi-level-recursion.md) - 动作:用动词
wait / observe / buy_probe / buy_confirmed / hold / reduce / sell_exit / rebuy / reject,附失效点+下一观察点。(仓位/风险见references/invalidation-risk.md) - 过滤器:结构归类完成后,才叠加 MACD/RSI/量/换手/资金流/趋势线/均线/筹码。(见
references/filters.md、references/volume-turnover-money.md) - 回测:要做策略/代码时,从一个最小闭环起步,严格定义与可测代理分列输出。回测硬数字与"风控非选股"的诚实口径见
references/empirical-evidence.md。(结构化代理规则见references/backtest-proxies.md) - 自检:新规则或重要结论,对照常见假阳性复核。(见
references/self-test-cases.md)
可视化请求
当用户要求“画出来 / 多周期标注 / 交互查看 / 导出分享”时,优先使用本仓库的本地工作台(见 references/visual-workbench.md),并遵守:
- 数据必须来自用户 CSV、内置合成示例或显式标注的可选公开行情源;不把合成数据当市场事实。
- 图上每个分型、笔、线段代理、中枢和背驰对象必须能回到原始 K 线范围及
confirmed_at / available_at。 - 多周期结果分层展示,不把 5m/30m 指标状态冒充日线结构确认。
- U1 多角度力度只作为可切换对照证据,不自动覆盖 MACD 基线或升级买卖点。
- 导出必须保留
definition_mode / approximation_loss / as_of / input_sha256 / execution_allowed=false。
工作台调用协议
- 先运行
chanlun-visual doctor --json;只把pass / pass_with_warnings当作可启动,不把命令存在等同于行情可用。 - 若找不到命令,说明运行时与 Skill 分开安装,给出
references/visual-workbench.md中的固定版本安装命令并等待用户授权;不得静默安装。 - 默认运行
chanlun-visual,只监听127.0.0.1:8791。除非用户明确要求并理解风险,不绑定局域网或公网地址。 - 优先使用内置合成示例或用户 CSV。可选公开行情失败时展示逐周期失败原因并回到 CSV,不把缺失历史补零。
- 若只需机器检查或启动,可运行
python scripts/skill_workbench.py check --json或python scripts/skill_workbench.py launch。
输出模板
**级别**
- trade_level / confirm_level / trigger_level / definition_mode / structure_completeness / approximation_loss
**结构**
- current_state / center / trend|oscillation / compared_movements
**信号**
- buy|sell class / divergence / confirmation / lower_level_trigger
**动作**
- action / basis / trigger / invalidation / next watch
何时加载哪个 reference
| 文件 | 用途 |
|---|---|
| concepts.md | 定义、源头层级、核心术语、状态分类 |
| strict-original-system.md | 原文自检门、严格/代理降级矩阵、编码契约 |
| structure-engine.md | K线包含、分型、笔、线段、中枢、延伸/扩张 |
| visual-reading.md | 图截图、书图、作者手绘图 |
| visual-workbench.md | 本地可视工作台、CSV 契约、图层、时点回放与导出 |
| buy-sell-playbooks.md | 一/二/三类买卖点流程、利润最大化模式、小转大 |
| multi-level-recursion.md | 区间套、小转大、同级别分解、大小级别冲突 |
| filters.md | MACD、RSI、趋势线、均线、筹码、板块强度 |
| volume-turnover-money.md | 量、换手、资金流的匹配规则 |
| invalidation-risk.md | 失效模式、止损、仓位、持有/减仓逻辑 |
| backtest-proxies.md | 把缠论变成可复现的代理规则 |
| empirical-evidence.md | 大样本回测硬数字:买点是试探非 alpha、卖点的风控价值、延迟敏感性 |
| self-test-cases.md | 常见假阳性、复核清单、可复用 case prompt |
怎么用(安装)
这是一个标准的 Agent Skill,分析规则与可视工作台运行时分开安装:
- Codex / OpenAI:从 GitHub 安装本 Skill,或把发布版 Skill ZIP 解压到
$HOME/.agents/skills/chanlun-trading-system/。需要可视界面时,再按references/visual-workbench.md安装chanlun-visual。 - Claude Code:把发布版 Skill 目录放进
~/.claude/skills/chanlun-trading-system/,新会话重新加载。 - 其他 Agent / 本地模型:加载
SKILL.md,需要细节时再按表读取对应references/*.md。 - 不会写代码的人:可使用仓库中的通用 AI Markdown;要获得完整交互界面,使用后续签名桌面安装包或由可信的人协助安装本地运行时。
来源与致谢
- 缠论(缠中说禅技术分析体系)原创归属 缠中说禅,本 Skill 是对其公开理论的操作化整理与研究性解读,用于学习与交流。
- 工作流设计参考了对原著《教你炒股票》及多本技术注解/图解/实战书的交叉精读,并以"严格原文 vs 可测代理"的分层来抑制过拟合。
- Skill 文件按仓库 MIT License 发布;缠论理论原始权利归其原作者。
Related Skills
last30days-skill
64.2kAI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
algorithmic-art
177.9kCreating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems.
pptx
177.9kUse this skill any time a .pptx or .potx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx or .potx file (even if the extracted content will be used elsewhere, like in an em…
design
133.6kComprehensive design skill: brand identity, design tokens, UI styling, logo generation (55 styles, Gemini, Atlas Cloud, or MuAPI AI), corporate identity program (50 deliverables, CIP mockups), HTML presentations (Chart.js), banner design (22 styles, social/ads/web/print), icon design (15 styles, SVG…
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
