FP&A prompts & agent repo
A cluster of posts shared AI prompt sets—'5 prompts that will change how you study accounting' and Perplexity-style prompts mimicking high-end advisors—plus a public GitHub repo of agent skills for financial analysis and trading. The content packages practical prompts and agent blueprints to convert raw financial data into crisp recommendations. (x.com) (x.com) (x.com)
A public repository titled "finance-skills-11" surfaces in the thread as a packaged collection of agent skills aimed specifically at financial analysis and trading, with a DEMOS.md and setup notes in its README. (github.com) The finance-skills-11 README includes an explicit “not financial advice” disclaimer and recommends installation as a Claude Code plugin, indicating the author intended the code for experimentation rather than production trading. (github.com) Parallel collections advertise “plug-and-play” skill modules that wire agents to real‑time news, price feeds, sentiment, and basic market-prediction utilities—RKiding’s Awesome-finance-skills describes exactly those capabilities in its repo blurb. (github.com) A separate research project, TradingAgents, documents a multi-agent framework that assigns roles (fundamental analyst, sentiment analyst, technical analyst, trader, risk manager) and reports experimental results on collaborative decision-making. (github.com) Examples of concrete modules across these repos include financial-data collectors that pull statements and metrics, automated ratio calculators (gross/operating/net margin, current ratio, debt/equity), risk tools like VaR and Sharpe, and templated recommendation generators that output buy/hold/sell guidance. (lobehub.com) Several of the published skill collections and indexes are explicitly labeled educational or research-grade and include demos/screenshots rather than production SLAs, signaling the need for FP&A-style validation, data governance, and mapping agent outputs to executive KPIs (revenue, margin, working capital) before use in C-suite decision documents. (github.com)
