The anthropics/financial-services repository is a Python-based collection of reference agents, skills, and data connectors designed for financial services workflows. It is classified across multiple fintech and banking domains including payment processing, financial APIs, account management, transaction monitoring, regulatory compliance, risk assessment, customer onboarding, and digital wallets. The repository shares contributors with major projects including microsoft/vscode, microsoft/typescript, and rust-lang/rust, suggesting cross-pollination with large open-source ecosystems.
The repository provides agents and skills for investment banking, equity research, private equity, and wealth management workflows. All functionality is available in two deployment modes: as Claude Cowork plugins or through the Claude Managed Agents API, allowing users to choose between integrated and headless deployment while maintaining identical system prompts and skills. The repository includes nine named agents covering distinct financial workflows. The Pitch Agent generates branded pitch decks end-to-end using comps and precedent analysis. The Market Researcher produces industry overviews and competitive landscapes from sector or theme inputs. The Earnings Reviewer processes earnings calls and filings to update models and draft notes. The Model Builder creates DCF, LBO, and three-statement models live in Excel. For fund administration, the Valuation Reviewer ingests GP packages and stages LP reporting, the GL Reconciler identifies and traces breaks, the Month-End Closer handles accruals and variance commentary, and the Statement Auditor reviews LP statements before distribution. The KYC Screener parses onboarding documents and runs rules engines to flag gaps.
The repository is organized into agents, vertical plugins, and skills. Vertical plugins bundle domain expertise by financial services segment: financial-analysis serves as the core with shared modeling skills and eleven data connectors, while investment-banking, equity-research, private-equity, wealth-management, fund-admin, and operations plugins add vertical-specific capabilities. Partner-built plugins for LSEG and S&P Global extend functionality with specialized data. The repository integrates with eleven external data providers via Model Context Protocol servers, including Daloopa, Morningstar, S&P Global, FactSet, Moody's, MT Newswires, Aiera, LSEG, PitchBook, Chronograph, and Egnyte, allowing Claude to access real-time market data, research, and document storage.
The repository includes tooling for deploying Claude inside Microsoft 365 applications against custom cloud infrastructure, enabling firms to run agents within Excel, PowerPoint, Word, and Outlook using their own Vertex AI, Bedrock, or internal LLM gateways. The repository is explicitly designed as a reference implementation for customization. Users can swap data connectors, inject firm-specific terminology and processes into skills, add branded templates, adjust agent scope to match team workflows, and extend with custom agents for uncovered workflows. All components are file-based using markdown and JSON with no build step required, making modification and deployment straightforward for financial services organizations.
