openclaw-data-china-stock - SKILL.md Agent Skill

name: openclaw-data-china-stock description: A-share/ETF/Option market data collection plugin for OpenClaw. tags: [data, china, etf, option, openclaw]

OpenClaw Data China Stock

This plugin provides a ClawHub/OpenClaw compatible toolset for collecting A-share, ETF, and listed option data.

Key tools

tool_fetch_market_data: Cross-asset unified entry (recommended).
tool_fetch_index_data / tool_fetch_etf_data / tool_fetch_option_data: Compatibility/alias unified entries.
tool_get_option_contracts: Fetch option contracts by underlying.
tool_read_market_data / tool_read_*: Read previously cached Parquet data (when enabled).

Paired skill suite

skills/china-macro-analyst/SKILL.md
skills/technical-analyst/SKILL.md
skills/market-scanner/SKILL.md
skills/market-sentinel/SKILL.md
skills/fund-flow-analyst/SKILL.md
skills/strategy-backtester/SKILL.md (MVP mode)
skills/fundamental-analyst/SKILL.md

China macro analyst tools

Primary: tool_fetch_macro_data / tool_fetch_macro_snapshot
Compatibility wrappers: tool_fetch_macro_* (21 tools; kept for backward compatibility)
Skill: skills/china-macro-analyst/SKILL.md (institutional 4-section narrative template; rules in macro_config.yaml)

Multi-factor equity screening (A-share)

tool_screen_equity_factors: Single entry for oscillation/trend templates (e.g. reversal_5d, fund_flow_3d, sector_momentum_5d) over hs300|zz500|zz1000|a_share|custom; returns quality_score, degraded, config_hash, optional sw_mapping stats. Implemented in plugins/analysis/equity_factor_screening.py; registered in tool_runner.py / config/tools_manifest.yaml (+ JSON). JSON Schema: docs/schemas/tool_screen_equity_factors.schema.json.
tool_batch_fetch: Batch runner includes tool_screen_equity_factors (see plugins/merged/tool_batch_fetch.py whitelist).
申万一级映射：config/sw_industry_level1_mapping.json，由 scripts/update_sw_industry_level1_mapping.py 生成（默认乐咕乐股 sw_index_first_info + sw_index_third_cons；可选 SW_MAP_USE_EM_SPOT=1 走东财快照）。消费侧：plugins/analysis/sw_industry_mapping.py。

助手侧夜盘落盘、质量门禁与熔断工具（tool_finalize_screening_nightly / tool_set_screening_emergency_pause）在 etf-options-ai-assistant 仓库注册；规程 Skill ota_equity_factor_screening_brief 仅在该助手仓维护并同步到 Gateway。

Fund-flow tools (pick one job, do not duplicate)

tool_capital_flow: Single-stock summary for workflows that need flow_judgement / risk_flags style outputs (e.g. limit-up strategies).
tool_fetch_a_share_fund_flow: Tabular / ranking / history for onshore A-share money flow (query_kind selects market/sector/stock tables, big deals, main-force ranks, sector drill-down). Uses Eastmoney/Tonghuashun-style AkShare routes with an explicit attempt chain; raw data is not investment advice.
tool_fetch_northbound_flow: Stock Connect northbound flows (cross-border), separate from onshore A-share flow tools—keep narratives and citations distinct.
tool_fetch_a_share_technical_screener: Tonghuashun-style technical stock screeners (new highs, consecutive up days, volume patterns, MA breakouts, etc.) via AkShare stock_rank_*_ths—not the same as locally computed MACD/RSI from OHLC (use tool_stock_data_fetcher / tool_calculate_technical_indicators for those).

Safety and independence

The plugin is designed to run independently (no dependency on any other repository):

It does not inject ~/.openclaw/.env.
It supports plugin-specific cache/data paths (configured inside the plugin).

Why it helps retail users

Unified cross-asset entry (tool_fetch_market_data) to reduce “tool switching” friction.
Multi-source provider priority + automatic fallback to avoid single-provider outages breaking your workflow.
Default read-only disk cache semantics (data_cache.enabled=false) to minimize local data pollution risk.

Typical usage

Example: fetch A-share index daily historical data:

tools:
  - name: tool_fetch_market_data
    params:
      asset_type: index
      view: historical
      asset_code: "000001"
      period: daily
      start_date: "20260201"
      end_date: "20260228"

Example: fetch ETF 5-minute bars:

tools:
  - name: tool_fetch_market_data
    params:
      asset_type: etf
      view: minute
      asset_code: "510300"
      period: "5"
      start_date: "20260201"
      end_date: "20260228"

Example: fetch latest CPI and use macro snapshot:

tools:
  - name: tool_fetch_macro_data
    params:
      dataset: cpi
      latest_only: true
  - name: tool_fetch_macro_snapshot
    params:
      scope: monthly
      include_quadrant: true

Cache policy

Default: data_cache.enabled=false
- Disk parquet reads are allowed (if cache exists).
- Disk parquet writes are skipped.
- Corrupted parquet files are not deleted in this mode.
data_cache.enabled=true
- Disk parquet reads and writes are both enabled.