finds.dev← search

// the find

himself65/finance-skills

★ 3,393 · JavaScript · MIT · updated Oct 2026

A collection of skills for AI financial analysis.

A library of Claude Code / Codex / Cursor "agent skills" for finance — stock valuation, earnings analysis, options payoff, technical strategies, plus read-only scrapers for Twitter, LinkedIn, Discord, Telegram, and TradingView. Aimed at someone who wants their coding agent to also do investment research and pull data from social/market sources without writing the glue code themselves.

Skills are organized into coherent plugin groups (market-analysis, social-readers, data-providers, startup-tools) rather than one flat pile, and most ship real reference material — WACC/ERP tables, Minervini's trend template, Amihud liquidity ratio — instead of just a one-line prompt wrapper. The social/data readers are explicitly read-only and lean on a shared scraping backend (opencli) instead of each skill reinventing auth and parsing. There's actual CI (skill-lint, opencli-plugin-test, release-skills) which is rare for a skills collection this size. The skill-creator plugin applying its own 10-dimension rubric to the other skills is a nice bit of eating your own dog food.

The valuation and strategy skills (DCF, SOTP, SEPA, ETF premium) are markdown instructions for an LLM to execute, not deterministic calculation code — correctness of the actual math rides entirely on the model getting WACC/DCF arithmetic right each run, with no unit-tested computation layer backing it. Several readers depend on scraping LinkedIn, Twitter/X, Discord, and TradingView's desktop app via CDP and cookie-based sessions through the external opencli project — brittle against site changes and shaky ground ToS-wise for at least LinkedIn. The dependency graph is large for what's conceptually a prompt library: yfinance, opencli, tdl for Telegram, headless TradingView MCP, CDP automation — each is its own failure point. Despite 'trading' as a topic, nothing here does backtesting or execution; it's all analysis framing, so the strategy skills are unverified against real market outcomes.

View on GitHub → Homepage ↗

// want more like this?

We dig through GitHub every week and send a few repos picked for what you actually care about — each with an honest take like this one.

Get finds in your inbox → Search again →