Ashare ST Risk Filter

• You are analyzing an A-share stock on the main board, ChiNext, or STAR Market and want to assess ST/*ST risk before or during a holding period.
• You want a structured, evidence-cited prediction after a quarterly earnings release (especially Q3 or interim reports) rather than a manual read of financial statements.
• You are monitoring a portfolio of A-share positions and need a repeatable, consistent risk-scoring process each earnings season.
• You are researching an already-ST company and want to assess whether it is likely to escalate to *ST or face forced delisting.

• The stock trades on the North Exchange (8xxxxx.BJ), Hong Kong, US markets, or is a crypto or commodity instrument — this skill explicitly rejects those.
• You need to detect financial fraud or predict the content of the next audit opinion — both are out of scope by design.
• You do not have a valid Tushare token; without it, all financial data falls back to akshare, which forces confidence down and may produce unreliable results.

Every analysis produces two independent ratings: a risk level (极高/高/中/低) and a prediction confidence level (高/中高/中/低). These are derived separately so a high-risk conclusion with low confidence is clearly distinguishable from a high-risk conclusion with high confidence.

The skill evaluates R1 (revenue + net profit, using min(n_income, profit_dedt)), R2 (year-end net assets), R3 (three-year cumulative dividend compliance), and R4 (consecutive loss chain with mandatory revenue linkage). Each uses board-specific thresholds for main board vs. ChiNext/STAR Market.

Penalties fetched from Sina Finance are scored over two time windows and weighted by subject type: company-level penalties count at 1.0×, while shareholder or officer penalties count at 0.5×. This prevents companies with shareholder-side violations from being systematically over-penalized.

When multiple data sources are available for the same metric, the skill enforces a strict priority order: earnings forecast or flash report > Q3 report > interim report > prior annual report. A higher-quality source always overrides a lower-quality one, and the source used is cited explicitly in the output.

All financial and fundamental data is fetched from tushare first. When tushare returns empty or fails due to insufficient credits, the skill automatically falls back to akshare and flags the downgrade in the output, reducing confidence by one level for any red line relying on akshare data.

When audit data is available, a single qualified or adverse opinion triggers a high-risk flag. When the fina_audit interface returns no data, the skill outputs direct URLs to cninfo, SSE, or SZSE disclosure pages so the user can verify manually — it never silently skips this check.

Before entering a position in an A-share stock, run this skill after the latest quarterly report to get a structured prediction of ST/*ST risk for the coming fiscal year, with explicit evidence citations.

After each Q3 or interim reporting period, re-run the skill across held positions to identify any stocks where financial deterioration has pushed one or more red lines into high-risk territory.

For stocks already carrying an ST designation, the skill shifts its analysis to predict *ST escalation or forced delisting probability, applying stricter thresholds appropriate to the current warning status.

Use the regulatory penalty dimension independently to identify companies with a high frequency of exchange warnings, regulatory letters, or inquiry letters — a leading indicator of 'compliance-type ST' even when financials look borderline.