himself65's finance-skills: 26 Agent Skills for Market Analysis and Research

Stock research means juggling price data, earnings estimates, sentiment, and short interest across a dozen tabs. finance-skills, a 3k-star collection from himself65, packages that workflow as 26 agent skills organized into six plugin groups that follow the Agent Skills open standard — your agent pulls the data, runs the frameworks, and shows the numbers.

Why This Skill Matters

The collection splits by job, and every skill maps to a source the README names:

• Market Analysis (11 skills) — company valuation with DCF, relative, and SOTP triangulation; earnings previews and recaps; analyst estimate deep-dives; ETF premium versus NAV; options payoff charts; SaaS valuation compression analysis; SEPA strategy with Minervini's trend template; stock correlation and liquidity; and yfinance-data, the market-data base layer for prices, financials, options, dividends, and earnings.
• Social Readers (6 skills) — read-only feeds from Twitter/X, Discord, LinkedIn, Telegram, and Y Combinator, plus a generic opencli fallback the README says covers 100+ other sources.
• Data Providers (6 skills) — sentiment via the Adanos Finance API, Fintel short interest and 13F ownership, Hormuz Strait monitoring, TradingView data through a bundled headless MCP server (plus a desktop-app reader sibling), and a Hyperliquid perp/spot reader.
• Startup Tools, UI Tools, and Skill Creator (1 each) — multi-perspective startup analysis, a generative UI system for rendering interactive widgets in Claude conversations, and a skill that evaluates other skills against a 10-dimension rubric.

Installation

The documented path is the plugin-groups installer, which prompts you to select one or more groups:

npx plugins add himself65/finance-skills

To install every group non-interactively, add --yes. To take a single skill instead, install it by name — yfinance-data is the natural first pick because the group's market data comes via yfinance:

npx skills add himself65/finance-skills --skill yfinance-data

For other agents, the README documents npx skills add himself65/finance-skills -a <agent-name>. Full documentation and demos also live at skills.himself65.com.

Real Workflow: A Pre-Earnings Briefing

Say you hold a stock reporting earnings this week. With the market-analysis group installed, brief your agent:

用已安装的 finance-skills 给我做一份财报前瞻简报：先拉这家公司近八个季度的价格和财务数据，然后生成财报前瞻（一致预期、超预期/不及预期历史、分析师情绪），最后算出它与同板块同行的相关性，标出适合配对交易的候选。

The agent runs yfinance-data for prices and financials, earnings-preview for the consensus-estimates, beat/miss history, and analyst-sentiment briefing, and stock-correlation for sector co-movement and pair-trading candidates. For a deeper pass afterwards, company-valuation triangulates an implied share price with WACC × g sensitivity and Bull/Base/Bear scenarios, and estimate-analysis shows whether the analysts revising estimates have historically been accurate.

Tips

• The README opens with its own warning: the project is for educational and informational purposes only, and nothing in it constitutes financial advice.
• The social readers are read-only by design — research feeds, not posting tools.
• tradingview-mcp runs a bundled headless MCP server, so TradingView data works with no app and no login; the sibling tradingview-reader covers what needs the desktop app, like full options chains with greeks.
• Install groups selectively — the installer's prompt lets you take only Market Analysis if the social and alternative-data feeds do not fit your workflow.
• generative-ui renders interactive HTML/SVG widgets inside the conversation, which turns a wall of numbers into readable output.

When Not to Use This

The README documents analysis and read-only data skills only, so you stay in your broker for anything that trades. Several data sources are third-party services with their own access requirements (Fintel, Adanos, TradingView, Hyperliquid). And if you need a one-off quote check, asking the agent to hit a free API directly is lighter than installing a 26-skill collection.

See the leaderboard for more skills.