About this skill

Problem it solves

Daily market reviews often become a manual assembly job: indices, sectors, limit-up activity, and northbound flow need to be checked and normalized every day. This skill targets that repetitive workflow. It does not require an API key or an account, and it focuses on A-shares, HK, US, Japan, Korea, Europe, and northbound flow, pulling public quotes locally to produce a shareable Markdown report.

How it works

The core flow is simple: scripts/daily_review.py emits structured data, and the template renders that data into a daily report. Each data section is fault-tolerant, so one failed feed is recorded in errors without stopping the whole run. The insights object only restates available facts: a summary, threshold-based alerts, and quality status such as loaded or missing sections. Rendering rules are strict: no buy or sell recommendations, no invented numbers, percentage values carry a sign, values keep two decimals, and unavailable sources are labeled plainly.

Boundaries

This is best for factual daily review, not stock picking or trade execution. It depends on public akshare data and local network access. Runs on weekends or holidays can fall back to the latest trading day, and if a feed times out or the dependency is outdated, some sections may be missing. In that case, trust the raw JSON and the reported gaps.