A-Share Daily Institutional News Digest

Problem

Daily A-share news feeds are noisy. Institutional views, policy signals, industry data, and company events are mixed together, making it hard to tell what is worth following. This skill frames the task as ingestion, multi-perspective filtering, and daily digest output, aiming to narrow “what happened today” into 20 readable, actionable items.

How It Works

The skill uses aggregated institutional messages from Zhishi Xingqiu and calls the zsrxq-cli-related API to fetch same-day posts. Because API responses may be truncated, the workflow expects repeated fetches when needed and filters out low-signal entries that contain only links and no body content.

Filtering uses three scoring passes:
- Macro-policy view: weighs impact breadth, policy importance, evidence reliability, and novelty.
- Trading view: evaluates whether the item can catalyze a short-term stock or sector move, with emphasis on direct catalyst strength, target clarity, and market attention.
- Information-asymmetry view: looks for underpriced, overlooked, or structurally important signals, emphasizing information gap and implied value.

Each round selects 20 candidates. The final list is ranked by frequency across the three rounds first, then by the highest single-round score as a tie-breaker, and outputs the top 20. The digest preserves key figures from the source, such as revenue, growth rate, price, and order volume, but does not expose the scoring process.

Boundaries

This skill is useful for reviewing same-day A-share institutional news, policy clues, and sector signals. It is a filtering workflow, not investment advice. The source emphasizes no guessing and no fabrication: judgments must be grounded in the actual content. Long industry summaries, routine broker “buy” notes, and old-news updates should not score high merely because they are lengthy or sound strong.