Giving LLM agents a Bloomberg terminal, piece by piece

What it does Awesome-finance-skills packages financial analysis capabilities into discrete agent “skills” — standardized folders containing SKILL.md files that drop into frameworks like OpenCode, Claude Code, or Antigravity. Instead of writing bespoke data pipelines, you get pre-wired modules for news aggregation, stock prices, sentiment scoring, and time-series forecasting that the agent can invoke directly.

The interesting bit The project treats financial analysis as an interoperability problem rather than a modeling one. By packaging everything from Polymarket odds to A-share fundamentals as uniform skill plugins — including a logic visualizer that spits out Draw.io XML — it effectively tries to build an npm-like registry for domain-specific agent capabilities.

Key highlights

• Eight skills covering news, stocks, sentiment, prediction, signal tracking, logic visualization, reporting, and search.
• News skill aggregates 10+ sources including Cailian, WSJ, Weibo, and Polymarket.
• Predictor uses a Kronos time-series model that adjusts forecasts based on news sentiment.
• Logic visualizer auto-generates transmission chain diagrams as Draw.io XML.
• Supports multiple agent frameworks via standardized SKILL.md folders.

Verdict Worth a look if you are assembling a financial analysis agent and would rather not spend your weekend maintaining scrapers. Skip it if you need a complete, end-to-end trading system — this is middleware, not a terminal.