AKShare#

Use this skill to fetch public financial and macro data with AKShare in a portable way.

Quick start#

1. Bootstrap the environment once#

Run:

bash

bash {baseDir}/scripts/bootstrap_akshare_env.sh

Default install target:

• $HOME/.openclaw/.venvs/akshare

Override it when needed:

bash

AKSHARE_VENV=/custom/path/.venv/akshare bash {baseDir}/scripts/bootstrap_akshare_env.sh

2. Run one-off queries with the helper#

bash

$HOME/.openclaw/.venvs/akshare/bin/python \
  {baseDir}/scripts/akshare_eval.py \
  --expr "ak.stock_zh_a_hist(symbol='000001', period='daily', start_date='20240101', end_date='20240131', adjust='')"

ak, pd, and json are preloaded in the expression.

Workflow#

1. Choose AKShare when it fits#

Prefer AKShare for:

• A股/港股/美股/ETF/基金行情
• 指数、板块、宏观、利率、债券、商品、期货等公开数据
• 快速拉表后做摘要、排序、统计、对比

If the user explicitly asks for another source such as Tushare, use that source instead.

2. Start narrow#

Fetch the smallest useful dataset first:

• narrow the symbol
• narrow the date range
• use the most specific AKShare function you can find

Avoid giant full-market pulls unless the user explicitly wants them.

3. Summarize instead of dumping raw tables#

After fetching:

• report the date range and the AKShare method used
• show only the key rows or aggregate values
• translate raw columns into plain Chinese when answering in Chinese

4. Handle upstream instability honestly#

AKShare wraps many public upstreams. If an endpoint fails:

• say the endpoint appears unstable or changed upstream
• try a nearby AKShare alternative if obvious
• do not present failed or partial results as authoritative

Common command pattern#

bash

$HOME/.openclaw/.venvs/akshare/bin/python \
  {baseDir}/scripts/akshare_eval.py \
  --expr "<AKShare expression>" --max-rows 20

Resources#

• Recipes: references/common-recipes.md
• Bootstrap installer: scripts/bootstrap_akshare_env.sh
• Query helper: scripts/akshare_eval.py

Maintainer / Updates#

• GitHub: https://github.com/Zack995
• X: https://x.com/btc_cczzc