stock-screener
Skill95 starsChanged 3 days agoMITView on GitHub
全市场股票筛选:按市场、估值、成长、趋势或自定义条件生成候选列表并排序。已知个股的深度分析用 stock-analysis。
12 other projects use a version of this.
Adopted by
Most recently taken up by
- Shubhamkumartrichy/claude-trading-skills-indiatheir own version, first seen Nov 15, 2026
- SentiSenseApp/skillstheir own version, first seen Nov 13, 2026
- Distilla-AI/distilla-skillstheir own version, first seen Nov 1, 2026
- ayushgoel/skillstheir own version, first seen Oct 13, 2026
- EodHistoricalData/eodhd-claude-skillstheir own version, first seen Oct 13, 2026
- rockchen1989/Stock-analyzer-agenttheir own version, first seen Oct 13, 2026
Used with
Projects with it also use technical-analysis.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
Bhala-SrinivashSkilltechnical-analysis
Systematic chart reading and technical analysis for NSE/BSE Indian equities. Covers trend identification, support/resistance mapping, volume analysis, indicator dashboards, and probabilistic scenario building. Uses Groww MCP for live data and yfinance for history.
Also in 3 projects that use this
Other versions12 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 12
- chengzuopeng/stock-sdk-mcp★ 102same name
- Shubhamkumartrichy/claude-trading-skills-india★ 64same name
- EodHistoricalData/eodhd-claude-skills★ 47same name
- Distilla-AI/distilla-skills★ 9same name
- kuntal-r-d/my-skills★ 3same name
- SentiSenseApp/skills★ 2same name
- ayushgoel/skills★ 1same name
- theepan/ai-agent-skills★ 1same name
- huangshiling40/claude-skills★ 0same name
- rockchen1989/Stock-analyzer-agent★ 0same name
2 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository21 other files it gives its agents.
- AGENTS.mdAGENTS.md
- bottom-volumeSkill
- box-oscillationSkill
- bull-trendSkill
- chan-theorySkill
- dragon-headSkill
- emotion-cycleSkill
- event-drivenSkill
- expectation-repricingSkill
- growth-qualitySkill
- hot-themeSkill
- ma-crossoverSkill
- market-reviewSkill
- one-yang-three-yinSkill
- sector-rotationSkill
- shrink-pullbackSkill
- stock-analysisSkill
- strategy-backtestSkill
- volume-breakoutSkill
- wave-theorySkill
- wisburg-researchSkill
The file
Files
全市场股票筛选
批量研究、子任务分工或报告文件交付时,必须读取宿主研究与报告交付,使用当前宿主实际提供的能力。
显式筛选条件缺少数据时工具返回 error,不能把该条件当作已通过。filters_applied 只包含实际执行的条件,基于返回的覆盖情况说明候选范围。
执行流程
第一步:确定筛选条件
优先从用户请求提取:
- 市场:A股 / 港股 / 美股(默认A股)
- 筛选偏好:价值型、成长型、趋势型、还是自定义条件
- 返回数量:默认 Top 20
未指定的项直接使用默认值并在结果中说明;只有自定义条件彼此冲突时才请用户补充。
第二步:运行筛选
运行数据采集脚本进行 L1 量化筛选:
python3 tools/screener.py screen --market A --top 20可选参数:--market(A/HK/US)、--top(返回数量)、--config(自定义 YAML 配置路径)、--l2(L2 量化增强开关,默认关闭)。
何时开启 --l2:需要对候选股做量化因子精排时开启。开启后逐股补算质量/成长/真实动量/波动率/资金流五类因子,结果更精细,但逐股抓取数据较慢,日常快速筛选保持关闭即可。开启 --l2 后候选按 scores.final 重新排序,config 里的 sort_by/sort_order 不生效;且 l2 模式下 scores.momentum 仍是 L1 参考值,不参与 final 综合分(final 用的是真实动量组)。
market_sentiment 输出字段:无论是否开 --l2,输出始终带 market_sentiment(市场情绪温度 0-100、level、multiplier 0.8-1.2 的 regime 调节器)。multiplier < 1 表示市场情绪偏弱,排序和推荐应更保守(收紧入选数量、降低激进度);multiplier > 1 表示情绪偏暖,可适度放宽。
第三步:L2 智能排序
对 L1 筛选出的候选股,基于返回的各项数据进行 LLM 智能排序。
注意区分两层"增强":--l2 是工具内的量化因子增强(在 L1 结果上补算因子,发生在排序之前);本步的 LLM 排序是 skill 流程的最后一环,在量化结果之上综合形态/量价/安全边际做最终排序。两者可叠加使用:先 --l2 拿精细因子,再做 LLM 排序。
排序考虑因素:
- 技术形态质量:趋势明确 > 震荡 > 趋势不明
- 量价配合度:放量上涨 > 缩量上涨 > 放量下跌
- 指标共振:多个指标同时看多/看空 > 指标分歧
- 安全边际:接近支撑 > 中间位置 > 接近压力
第四步:输出推荐列表
## 选股结果 — [市场] Top [N]
### 筛选条件
- 市场:...
- 过滤条件:...
- 全市场 → L1 筛选 → L2 排序
| 排名 | 代码 | 名称 | 价格 | 涨跌幅 | PE | PB | 综合评分 | 推荐理由 |
|------|------|------|------|--------|----|----|----------|----------|
| 1 | ... | ... | ... | ... | .. | .. | ... | ... |
### 重点关注
对排名前3的股票给出简要分析
### 风险提示
- 筛选基于历史数据,不代表未来表现
- 建议结合个股深度分析后再做决策每个候选必须读取 factor_coverage.available/missing/ratio/effective_weights。ratio 是已覆盖因子的原始权重之和;L1 与 L2 的 effective_weights 均按可用因子重归一化,缺失或非有限值不按零分计,排序前先按实际可用因子计算综合分,effective_scores 保留实际参与计算的各组分数;用 sum(effective_scores[g] × effective_weights[g]) × market_sentiment.multiplier 核对 scores.final(四舍五入至一位)。scores.momentum 仍保留 L1 参考值,L2 贡献以 effective_scores.momentum 为准。相同分数但缺失因子不同的候选不能视为同等证据,报告缺失项和覆盖度,不以 enriched=true 代替覆盖判断。scores.final=null 表示没有足够因子计算,不能作为高分候选推荐。
估值比较还需读取 pe_basis:dynamic、trailing 与 unknown 不得当成同一 PE 口径。筛选中的有限非正 PE/PB 价值子分为 0,不能因数值更低而获得最高价值分;缺失或非有限因子仍保持未知,不按零分处理。
L2 的 vol20 只使用恰好 20 个有效日收益,factor_coverage.vol20_samples 记录实际样本数,非正/非有限收盘价先剔除,空输入为 0;短样本为缺测。质量因子按 report_date 分组归一,年度累计与单季累计不得直接排名;资金流 5/10/20 日窗口不足时返回 null,不按 0 或完整窗口处理。
完成标准:列出实际生效的市场、过滤条件、排序口径和候选数量,且推荐理由只使用返回数据,并说明因子覆盖。
策略模板
如果用户没有明确条件,提供预设策略:
- 价值型:低PE + 低PB + 高ROE + 稳定分红
- 成长型:中等PE + 高营收增长 + 高换手率
- 趋势型:放量突破 + 均线多头 + 高量比
- 低估反转:超跌 + 低PB + 资金净流入
股息因子(A股,可选)
配置 filters.dividend_yield_min / dividend_yield_max 或 sort_by: dividend_score / dividend_yield 时才拉取股息数据。dividend_report_date 可指定报告期,默认上一年年末。这里使用已实施方案的报告期股息率(百分数),不是滚动十二个月股息率;缺失为未知,使用股息阈值时排除。启用后 value 分项加入股息横截面评分,未启用时不增加股息请求。
filters:
dividend_yield_min: 3
sort_by: dividend_score
sort_order: desc
dividend_report_date: "2025-12-31"筛选结果仍需用 get_dividends(mode="history", symbol=...) 核查已实施现金分红、连续支付年数与除息时间,不能把高股息直接解释为低风险。
补充数据覆盖
股息筛选支持 A/HK:A 股为报告期已实施方案股息率,HK 为 HKEX/LSEG 年度股息率。两者都不是当前价 TTM;使用结果的 yield_basis 和 report_date,缺失不当零。
选股的 top 必须为正整数;超时预算按候选数量与采集批次计算。
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

