AI Agent Hub
Back to skills
RockSolid Stock Rating Model icon

RockSolid Stock Rating Model

Professional Updated 2026.08.29

Paste the following prompt into your AI chat to install this skill:

Please follow https://skillhub.cn/install/skillhub.md to install @user_525a1b37/rocksolid-stock-rating.

About this skill

Problem to solve

Stock screening often mixes single-metric judgments, market noise, and inconsistent data sources. RockSolid organizes quotes, financial statements, valuation, and risk signals into an executable rating workflow for A-share fundamental diagnostics.

How it works

The skill follows the execution manual in SKILL.md, then loads scripts and references only when needed. Key capabilities include:
- Data fetching: prefers AKShare, falls back to Sina, East Money, and other HTTP endpoints, and can use Agent-driven TDX MCP to fill missing fields.
- Scoring: rock_solid.py calculates cluster scores, threshold checks, adjustments, and final scores instead of manual estimation.
- Trend analysis: trend_analyzer.py compares multi-period financial reports using matching reporting periods to track fundamental score evolution.
- Output: the --report mode shows data-source status, missing-field warnings, scoring details, and formulas for better auditability.

Boundaries

It mainly supports the Shanghai and Shenzhen main boards, ChiNext, and the STAR Market. The Beijing Stock Exchange and B-shares may have incomplete data and lower confidence. Funds, ETFs, convertible bonds, HK stocks, and US stocks are rejected or unsupported. It is an analytical aid based on public historical data and preset rules, not investment advice.

Use Cases

  • Given A-share codes such as 600519, run a rule-based rating and inspect data-source status, threshold checks, and final score.
  • Suspecting blowup risk, review financial statements, cash flow, debt, and pledge fields while checking missing-field warnings.
  • Comparing two stocks over two years, fetch multi-period reports by matching reporting periods to track fundamental score trends.
  • When Python or data endpoints are unstable, backfill quotes and financials via HTTP or TDX MCP before scoring instead of estimating.

Best For

  • Quant researchers tracking A-share fundamentals who need reproducible score conclusions from financials, valuation, and quotes.
  • Sell-side or buy-side analysts preparing stock research who need to audit data sources, default fields, and adjustments.
  • Financial Agent engineers integrating rating scripts, data fallback chains, and TDX MCP data flows.
  • Risk-control operators screening for blowup risk who need quick public-field checks on cash flow, debt, and earnings quality.