Claude’s Open‑Source Financial Skills: A Deep Dive

Anthropic’s new claude‑for‑financial‑services repository bundles 11 ready‑to‑run agents, vertical plugins, and 11 MCP data connectors that automate core Wall Street workflows—from pitch decks and earnings reviews to valuation modeling—while offering clear installation paths and guidance for enterprise customization.

Old Zhang's AI Learning
Old Zhang's AI Learning
Old Zhang's AI Learning
Claude’s Open‑Source Financial Skills: A Deep Dive

1. What It Is

Anthropic has released the claude-for-financial-services repository, an end‑to‑end plugin package that reproduces the daily work of Wall Street analysts. The agents draft work papers (models, memos, research reports, reconciliations) but do not make investment decisions, execute trades, assume risk, or open accounts, keeping responsibility clearly separated.

2. The 11 Agents and Their Scenarios

Pitch Agent : Generates a branded pitch deck with comparable companies, precedent transactions, and LBO analysis.

Meeting Prep Agent : Produces a briefing pack automatically before client meetings.

Market Researcher : Given a sector or theme, returns an industry overview, competitive landscape, peer comps, and a target list.

Earnings Reviewer : Processes earnings call transcripts and announcements, updates models, and drafts research reports.

Model Builder : Runs DCF, LBO, three‑statement, and comparable‑company models directly inside Excel.

Valuation Reviewer : Consumes GP reporting packages, runs valuation templates, and prepares LP reports.

GL Reconciler : Finds general‑ledger breaks, traces origins, and follows sign‑off procedures.

Month‑End Closer : Handles month‑end accruals, roll‑forwards, and variance explanations.

Statement Auditor : Audits LP statements before distribution.

KYC Screener : Parses onboarding documents, runs rule engines, and flags gaps.

Additional vertical plugins (7 industry packs + 2 partner packs) provide the underlying skills, slash commands, and data connectors for specific financial sub‑domains.

3. Core Skills (the Real Treasure)

The vertical plugins are built on the financial-analysis core package, which bundles shared modeling skills and eleven data‑connector plugins. Notable skills include: /comps: Comparable‑company analysis with transaction multiples. /dcf: Discounted cash‑flow valuation with WACC and sensitivity analysis. /lbo: Leveraged‑buyout modeling. /3‑statement‑model: Three‑statement model population. /debug‑model (audit‑xls): Excel model audit for formula tracing, hard‑code detection, and balance checks. /ppt‑template: Teaches Claude a company‑specific PowerPoint template.

4. The 11 MCP Data Connectors (the Real Moat)

Anthropic provides built‑in adapters for premium financial data sources, including Daloopa, Morningstar, S&P Global/Capital IQ, FactSet, Moody’s, MT Newswires, Aiera, LSEG, PitchBook, Chronograph, and Egnyte. Access to many of these connectors requires a subscription or API key from the data provider.

5. Installation: Three‑Line Claude Code Commands

For Claude Cowork users, add the marketplace entry anthropics/claude-for-financial-services and install the core package and desired agents via the UI. For command‑line fans, the recommended steps are:

# 1. Add marketplace
claude plugin marketplace add anthropics/claude-for-financial-services

# 2. Install core package (includes all data connectors)
claude plugin install financial-analysis@claude-for-financial-services

# 3. Install selected agents
claude plugin install pitch-agent@claude-for-financial-services
claude plugin install gl-reconciler@claude-for-financial-services
claude plugin install market-researcher@claude-for-financial-services

# 4. Install vertical industry packs as needed
claude plugin install investment-banking@claude-for-financial-services
claude plugin install equity-research@claude-for-financial-services

Agents appear in the Cowork dispatch list, and slash commands such as /comps, /dcf, /earnings, and /ic‑memo can be invoked directly. For self‑hosted deployments, use the Managed Agents API ( /v1/agents) with a script that uploads skills and registers leaf‑worker sub‑agents.

6. Real Positioning

The repository is not a finished product but a reference implementation. The README stresses that the templates are meant to be adapted to a company’s own data sources, terminology, and workflow boundaries. Five recommended customizations are:

Replace the .mcp.json file to point to internal data providers.

Inject company‑specific terminology and formats into skill files.

Provide your own PowerPoint templates via /ppt‑template.

Adjust agent boundaries by editing agents/<slug>.md.

Copy the existing structure to add new agents or skills.

The repo is file‑based (Markdown + YAML) with no build step, making it highly extensible for internal IT teams.

7. Who Should Read This

Investment bankers, sell‑side research analysts, PE/VC professionals, family‑office staff, and fund‑operations teams who recognize the listed skills as core to their daily work.

AI solution providers serving financial institutions – the repo serves as a ready‑made RFP response.

Developers learning how to write production‑grade Claude Skills – the official implementation is the best reference.

8. Author’s Judgment

Anthropic’s move is not about selling subscriptions; it aims to set a standard for enterprise‑grade AI agents. The architecture is likely to be replicated in healthcare, legal, consulting, and government domains. For ordinary developers, the repo is a free, high‑quality textbook on building usable Claude Skills.

Original Source

Signed-in readers can open the original source through BestHub's protected redirect.

Sign in to view source
Republication Notice

This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactadmin@besthub.devand we will review it promptly.

AI agentsopen sourceInstallationClaudeFinancial ServicesSkillsMCP Connectors
Old Zhang's AI Learning
Written by

Old Zhang's AI Learning

AI practitioner specializing in large-model evaluation and on-premise deployment, agents, AI programming, Vibe Coding, general AI, and broader tech trends, with daily original technical articles.

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.