crypto-regime-analyzer

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Crypto Regime Analyzer Skill

Crypto Regime Analyzer 技能

Purpose

用途

Quantify the crypto market regime using a data-driven 6-component scoring system (0-100). This is the crypto analog of

market-breadth-analyzer

exposure-coach

: it answers "what posture does the crypto market currently support?" before any coin-level analysis happens.

Score direction: 100 = Maximum risk-on health (broad participation, healthy trend, sane leverage), 0 = Critical risk-off.

No API key required — uses CoinGecko's free public API and Binance's public futures endpoint.

采用数据驱动的6维度评分体系（0-100分）量化加密货币市场状态。这是加密货币领域对标

market-breadth-analyzer

exposure-coach

的工具：在进行任何币种层面的分析之前，它可以回答“当前加密货币市场适合何种仓位策略？”的问题。

评分说明： 100分=风险偏好型市场健康度拉满（广泛参与、趋势健康、杠杆合理），0分=极端风险规避型市场。

无需API密钥 —— 使用CoinGecko的免费公开API和Binance的公开期货端点。

When to Use This Skill

适用场景

User asks "Is crypto risk-on or risk-off right now?" or "How healthy is the crypto market?"
User asks "Is it alt season?" or about BTC dominance direction
User asks whether funding rates are overheated
User wants an exposure posture for a crypto sleeve before screening individual coins
User wants a daily crypto regime check alongside the equity
```
market-regime-daily
```
workflow

用户询问“当前加密货币市场是风险偏好还是风险规避状态？”或“加密货币市场健康度如何？”
用户询问“是否处于山寨币季节？”或BTC主导地位的变化方向
用户询问资金费率是否过热
用户在筛选单个币种之前，希望为加密货币资产配置获取仓位建议
用户希望将每日加密货币状态检查与股票
```
market-regime-daily
```
工作流结合使用

What This Skill Does NOT Do

不支持的功能

No coin picks, no buy/sell signals, no price targets
No execution or portfolio changes — regime description only
Human decision gates remain central, consistent with the project vision

不提供币种推荐、买卖信号或价格目标
不执行交易或调整投资组合——仅描述市场状态
人工决策仍是核心，与项目愿景一致

Prerequisites

前置条件

Python 3.9+ with
```
requests
```
(live mode only; offline mode is stdlib-only)
Internet access to
```
api.coingecko.com
```
and
```
fapi.binance.com
```
(live mode)
No API keys required

Python 3.9+ 环境，需安装
```
requests
```
库（仅在线模式需要；离线模式仅依赖标准库）
网络访问权限 可连接
```
api.coingecko.com
```
和
```
fapi.binance.com
```
（在线模式）
无需API密钥

Component Model

维度模型

#	Component	Weight	Question it answers
1	BTC Trend Structure	25%	Is the reserve asset's primary trend intact? (price vs 50/200DMA stack, 200DMA slope)
2	Alt Breadth Participation	20%	How broadly are alts participating? (% of top-N above 200DMA, 50DMA confirmation)
3	BTC Dominance Regime	15%	Where is capital rotating? (dominance direction interpreted jointly with BTC trend)
4	Perpetual Funding Regime	15%	How crowded is leverage? (avg funding across majors; contrarian at extremes)
5	Drawdown & Volatility Position	15%	Where are we in the cycle? (drawdown from 1y high, realized vol percentile)
6	Momentum Thrust / Washout	10%	Short-horizon confirmation (% of universe positive over 30d)

Missing components have their weight proportionally redistributed (same convention as

market-breadth-analyzer

). Full scoring logic:

references/crypto_regime_methodology.md

#	维度	权重	解决的问题
1	BTC趋势结构	25%	储备资产的主要趋势是否完好？（价格与50/200DMA均线对比、200DMA均线斜率）
2	山寨币广度参与度	20%	山寨币的参与广度如何？（前N大币种中价格高于200DMA的占比、50DMA确认情况）
3	BTC主导地位状态	15%	资金流向何方？（结合BTC趋势解读主导地位变化方向）
4	永续合约资金费率状态	15%	杠杆持仓拥挤程度如何？（主流币种平均资金费率；极端值为反向指标）
5	回撤与波动位置	15%	我们处于周期的哪个阶段？（距1年高点的回撤幅度、已实现波动率百分位）
6	动量推力/洗盘	10%	短期趋势确认（30天内全市场币种上涨占比）

缺失维度的权重会按比例重新分配（与

market-breadth-analyzer

遵循相同规则）。完整评分逻辑见：

references/crypto_regime_methodology.md

。

Regime Zones

状态区间

Score	Zone	Posture
80-100	RISK_ON	Broad risk-on conditions observed; review risk limits before decisions
40-79	NEUTRAL	Mixed conditions observed; no strong regime conclusion
0-39	RISK_OFF	Defensive market conditions observed; review existing risk controls

These are heuristic descriptive bands, not validated allocation rules. See

references/VALIDATION.md

for the current evidence boundary and the artifacts required before making quantitative performance claims.

评分	区间	仓位建议
80-100	RISK_ON	观察到广泛的风险偏好型市场条件；决策前需审查风险限额
40-79	NEUTRAL	观察到混合市场条件；无法得出明确的市场状态结论
0-39	RISK_OFF	观察到防御型市场条件；需审查现有风险控制措施

这些是启发式描述区间，并非经过验证的配置规则。有关当前证据边界以及提出量化性能声明所需的文档，请参阅

references/VALIDATION.md

。

Execution Workflow

执行流程

Phase 1: Run the Analysis Script

阶段1：运行分析脚本

Live mode (fetches CoinGecko + Binance; first run of the day takes ~2-4 minutes at the default

--top-n 20

due to free-tier rate-limit throttling; same-day re-runs hit the cache and are instant):

bash

mkdir -p reports/<routine-or-date>
python3 skills/crypto-regime-analyzer/scripts/crypto_regime_analyzer.py \
  --output-dir reports/<routine-or-date>

Offline mode (no network; snapshot schema in the methodology reference):

bash

python3 skills/crypto-regime-analyzer/scripts/crypto_regime_analyzer.py \
  --input-json snapshot.json \
  --output-dir reports/<routine-or-date>

Options:

--top-n <int>

universe size (default 20),

--cache-dir <path>

fetch cache location (default

.crypto_regime_cache

--quiet

在线模式（获取CoinGecko + Binance数据；默认

--top-n 20

参数下，每日首次运行需约2-4分钟，因免费版速率限制；当日重新运行会命中缓存，瞬间完成）：

bash

mkdir -p reports/<routine-or-date>
python3 skills/crypto-regime-analyzer/scripts/crypto_regime_analyzer.py \
  --output-dir reports/<routine-or-date>

离线模式（无需网络；快照 schema 见方法论参考文档）：

bash

python3 skills/crypto-regime-analyzer/scripts/crypto_regime_analyzer.py \
  --input-json snapshot.json \
  --output-dir reports/<routine-or-date>

可选参数：

--top-n <int>

市场币种范围（默认20），

--cache-dir <path>

缓存存储路径（默认

.crypto_regime_cache

），

--quiet

静默模式。

Phase 2: Interpret the Output

阶段2：解读输出结果

The script writes

crypto_regime.json

(machine-readable, for chaining into other skills) and

crypto_regime.md

(one-page report), and prints a one-line summary:

CRYPTO REGIME: NEUTRAL (score 68.4/100) — Mixed conditions observed; no strong regime conclusion

When presenting results, lead with the zone and posture, then explain the 1-2 components most responsible for the score using their

signal

strings. Flag any components reporting

data_available: false

and what that means for confidence.

脚本会生成

crypto_regime.json

（机器可读格式，可用于与其他技能联动）和

crypto_regime.md

（单页报告），并打印一行摘要：

CRYPTO REGIME: NEUTRAL (score 68.4/100) — Mixed conditions observed; no strong regime conclusion

展示结果时，先说明区间和仓位建议，再解释对评分影响最大的1-2个维度的

signal

内容。标记任何显示

data_available: false

的维度，并说明其对结果可信度的影响。

Phase 3 (optional): Feed Downstream

阶段3（可选）：下游联动

The JSON composite can slot into an

exposure-coach

-style posture summary as one descriptive crypto-market input. It must not independently authorize, block, size, or execute a trade.

JSON格式的综合结果可作为描述性加密货币市场输入，纳入

exposure-coach

风格的仓位摘要中。它不得独立授权、阻止、调整或执行交易。

Output

输出内容

The script writes two artifacts to

--output-dir

and prints a one-line summary for workflow chaining:

```
crypto_regime.json
```
— full machine-readable analysis:
```
metadata
```
, per-component results (
```
score
```
,
```
signal
```
,
```
data_available
```
, component-specific fields), and the
```
composite
```
block (
```
score
```
,
```
zone
```
,
```
guidance
```
,
```
effective_weights
```
).
```
crypto_regime.md
```
— one-page report: composite score with zone bar, posture line, per-component table (weight / score / signal), and confidence notes.

Console:

CRYPTO REGIME: <ZONE> (score <N>/100) — <posture>

plus warnings for any skipped components.

脚本会向

--output-dir

写入两个文件，并打印一行摘要用于工作流联动：

```
crypto_regime.json
```
—— 完整的机器可读分析结果：包含
```
metadata
```
、各维度结果（
```
score
```
、
```
signal
```
、
```
data_available
```
、维度特定字段），以及
```
composite
```
模块（
```
score
```
、
```
zone
```
、
```
guidance
```
、
```
effective_weights
```
）。
```
crypto_regime.md
```
—— 单页报告：包含带区间进度条的综合评分、仓位建议、各维度表格（权重/评分/信号），以及可信度说明。
控制台输出：
```
CRYPTO REGIME: <ZONE> (score <N>/100) — <posture>
```
以及任何跳过维度的警告信息。

Resources

资源

```
references/VALIDATION.md
```
— validation status, evidence boundary, and reproduction requirements.
```
references/crypto_regime_methodology.md
```
— full scoring rationale, every threshold table, the offline snapshot JSON schema, and the live data-source endpoint list.
```
scripts/crypto_regime_analyzer.py
```
— CLI orchestrator (entry point).
```
scripts/data_client.py
```
— CoinGecko/Binance fetchers, per-day cache, dominance history accumulator, offline loader.
```
scripts/calculators/
```
— one module per component; pure functions, fully unit-tested.
```
scripts/scorer.py
```
— weighted composite with proportional weight redistribution.
```
scripts/tests/
```
— tests covering every component, the scorer, sparse-data fail-closed behavior, and end-to-end bull/bear/degraded snapshots.

```
references/VALIDATION.md
```
—— 验证状态、证据边界和复现要求。
```
references/crypto_regime_methodology.md
```
—— 完整评分依据、所有阈值表、离线快照JSON schema，以及在线数据源端点列表。
```
scripts/crypto_regime_analyzer.py
```
—— CLI协调器（入口文件）。
```
scripts/data_client.py
```
—— CoinGecko/Binance数据获取器、每日缓存、主导地位历史累加器、离线加载器。
```
scripts/calculators/
```
—— 每个维度对应一个模块；纯函数，已完成单元测试。
```
scripts/scorer.py
```
—— 带权重比例重分配的加权综合评分器。
```
scripts/tests/
```
—— 涵盖所有维度、评分器、稀疏数据故障关闭行为，以及牛/熊/降级快照端到端测试的测试用例。

Known Limitations

已知限制

Dominance history accumulates locally. CoinGecko's free tier only exposes current dominance, so the client stores one observation per run-day in the cache dir. The dominance component reports
```
data_available: false
```
until 31 daily observations exist (weight is redistributed until then). Seed it faster via
```
--input-json
```
.
Funding is best-effort. If Binance's endpoint is unreachable (geo-restrictions, outage), the component is skipped gracefully.
Universe is top-N by market cap with stablecoins and wrapped/staked assets excluded; it is not a fixed index, so composition drifts with the market.
Thresholds are heuristic and documented in the methodology reference; they are conservative defaults, not backtested optima.

主导地位历史数据本地累加。CoinGecko免费版仅提供当前主导地位数据，因此客户端会在缓存目录中存储每日运行的一次观测数据。主导地位维度在积累31条每日观测数据之前会显示
```
data_available: false
```
（在此期间权重会被重新分配）。可通过
```
--input-json
```
参数快速填充数据。
资金费率数据尽最大努力获取。若Binance端点无法访问（地域限制、服务中断），该维度会被优雅跳过。
币种范围为市值前N大币种，排除稳定币和包装/质押资产；这并非固定指数，因此组成会随市场变化而漂移。
阈值为启发式设定，已记录在方法论参考文档中；它们是保守默认值，并非经过回测的最优值。

Disclaimer

免责声明

Educational and process-improvement use only. This skill describes market conditions; it does not provide financial advice, signals, or buy/sell instructions. All decisions remain the user's responsibility.

仅用于教育和流程改进。本技能仅描述市场状况；不提供财务建议、交易信号或买卖指令。所有决策由用户自行负责。