Plug-in analyst for Claude Code: DCF, sentiment, and Hormuz Strait
A modular skill pack that turns an AI agent into a financial research assistant with market data, social feeds, and even geopolitical risk monitors.

What it does
finance-skills is a collection of agent plugins built for the Agent Skills open standard, installable via npx into Claude Code or other agents. The modules cover stock valuation, earnings analysis, social-media monitoring, sentiment tracking, and a few oddities like Strait of Hormuz shipping alerts. Most market data flows through yfinance; social and third-party feeds rely on the opencli adapter ecosystem.
The interesting bit
The project treats financial analysis as a composable skill layer rather than a monolithic app. A user can install just the options-payoff charting skill, or the full finance-market-analysis plugin, or mix in twitter-reader and tradingview-reader to build a personalized research bot. The skill-creator module even scores new skills against a 10-dimension rubric, which suggests the author expects this catalog to keep growing.
Key highlights
- Market analysis: DCF triangulation, earnings previews/recaps, ETF premium/discount, SEPA trend strategy, liquidity metrics, and interactive options payoff charts.
- Social readers: Read-only feeds for Twitter/X, Discord, LinkedIn, Telegram, Y Combinator, plus a generic fallback covering 90+ sources via
opencli. - Data providers: Sentiment from Adanos (Reddit, X, news, Polymarket), analyst-grade synthesis from Funda AI, and a TradingView desktop reader that scrapes quotes and options greeks via CDP.
- Geopolitical oddity: A dedicated Hormuz Strait monitor for shipping, oil impact, and insurance risk — not your typical fintech feature.
- Generative UI: HTML/SVG widget rendering inside Claude conversations via
show_widget.
Caveats
- The README carries a prominent “educational and informational purposes only” warning; this is not a trading execution system.
- Several skills are thin wrappers around external APIs (
yfinance,opencli, Funda AI, Adanos) — value lies in the curation and prompt engineering, not in novel data pipelines. - No benchmarks, latency figures, or accuracy metrics are provided for the valuation or sentiment skills.
Verdict
Worth a look if you already use Claude Code and want pre-built financial research prompts without wiring APIs yourself. Skip it if you need live trading execution, audited valuation models, or a standalone application — this is glue code for agent conversations, and it knows it.