anthropics/financial-services in Claude Code: agents for finance teams
By Flowy · Updated 2026-08-28
anthropics/financial-services is Anthropic's reference collection of Claude agents, skills, commands, and connectors for financial services work: investment banking, equity research, private equity, wealth management, fund administration, and KYC operations. It packages 10 agents built from 67 skills and 56 commands, plus LSEG and S&P Global data connectors, drafting analyst work product for human review.
What is the anthropics/financial-services plugin, actually
anthropics/financial-services is Anthropic's own reference repository for using Claude across financial services work: investment banking, equity research, private equity, and wealth management, plus fund administration and client onboarding. Flowy indexes it as one listing of 135 components: 67 skills, 10 agents, 56 commands, and 2 MCP connectors.
The repository is explicit about what these components are not. Its own documentation states that the agents draft analyst work product, models, memos, research notes, reconciliations, for a qualified professional to review. They do not make investment recommendations, execute transactions, bind risk, post to a ledger, or approve onboarding. Every output is staged for a human to sign off before it goes anywhere.
What is inside
The 10 agents are named, end to end entry points, and each one bundles every skill it needs so installing the agent is enough on its own. The 67 skills and 56 commands underneath them are organized by financial services vertical rather than handed out loose, and the 2 MCP connectors wire two of those verticals to a partner's live market data.
The ten agents
Pitch Agent turns comps, precedents, and an LBO into a branded pitch deck. Market Researcher turns a sector or theme into an industry overview, a competitor map, and an ideas shortlist. Earnings Reviewer takes an earnings call and filings through to a model update and a note draft. Model Builder runs DCF, LBO, and 3-statement work live in Excel, and Meeting Prep Agent builds a briefing pack before a client meeting.
On the finance ops side, GL Reconciler finds breaks and traces the root cause, Month-End Closer handles accruals, roll-forwards, and variance commentary, Statement Auditor audits LP statements before distribution, and Valuation Reviewer stages LP reporting from GP packages. KYC Screener parses onboarding documents against a rules engine and flags gaps.
The verticals underneath them
The shared modeling core covers comps-analysis, dcf-model, lbo-model, and 3-statement-model, alongside audit-xls for formula tracing and hardcode detection in an existing workbook. That core pairs a skill with a command of the same shape, so comps-analysis fires as comps and dcf-model fires as dcf.
Deal work adds cim-builder, teaser, buyer-list, merger-model, and deal-tracker. Coverage and publishing add earnings-analysis, initiating-coverage, thesis-tracker, and catalyst-calendar. Private equity adds deal-sourcing, dd-checklist, unit-economics, ic-memo, and value-creation-plan. Wealth management adds client-review, financial-plan, portfolio-rebalance, and tax-loss-harvesting. Fund administration adds gl-recon, break-trace, accrual-schedule, and nav-tieout. Onboarding adds kyc-doc-parse and kyc-rules. A few skills, like xlsx-author and skill-creator, are meant to run automatically rather than be typed as a command at all.
The two MCP connectors: lseg and spglobal
lseg wires Claude to LSEG's analytics API behind fixed income and macro skills like bond-relative-value, swap-curve-strategy, fx-carry-trade, and macro-rates-monitor. spglobal wires Claude to S&P's Kensho data platform behind tear-sheet, earnings-preview-beta, and funding-digest. Both are partner-built additions layered on top of the shared modeling core, and the core plugin's own connector file documents a longer list of optional data providers beyond these two.
What the verify skill actually checks
One skill here is easy to misread. verify is not a financial-modeling check, it belongs to a separate part of the repository: admin tooling for provisioning Claude inside Excel, PowerPoint, Word, and Outlook through the Microsoft 365 add-in. It is written for someone changing that tool's own install scripts, not for an analyst, and it drives cache-clearing scripts against a sandboxed home directory to confirm that removing a stale add-in manifest never touches a user's chat history. For checking an actual financial model, the relevant skill is audit-xls, run through the debug-model command.
How do I install it
Install from the listing page linked at the top of this guide. Flowy indexes what anthropics/financial-services ships. It does not own, bundle, or modify the repository: the agents, skills, and connectors stay on their author's releases in the source repo, and updates flow from there, not from Flowy. The project's own instructions are direct about the rest of the deal: tune the prompts and connectors to your firm's conventions, and keep a qualified professional in the review loop, because nothing here is built to skip that step.
Common questions
- Does the anthropics/financial-services plugin make investment decisions on its own?
- No. The repository's own documentation states that these agents draft analyst work product, models, memos, research notes, and reconciliations, for a qualified professional to review. They do not make investment recommendations, execute transactions, bind risk, post to a ledger, or approve onboarding. Every output is staged for a human to sign off before it is used.
- What is the difference between an agent and a vertical plugin in this repository?
- An agent, such as Pitch Agent or GL Reconciler, is a self-contained bundle built around one named workflow, and it ships its own copy of every skill it needs, so installing it is enough on its own. A vertical plugin, such as the investment-banking or equity-research bundle, is the underlying set of skills and commands for a whole area of financial services work, meant to be installed on its own if you just want the commands and connectors without a full agent.
- What do the LSEG and S&P Global connectors actually add?
- They are the two MCP connectors this listing indexes. The lseg connector wires Claude to LSEG's analytics API behind skills like bond-relative-value, swap-curve-strategy, and macro-rates-monitor. The spglobal connector wires Claude to S&P's Kensho data platform behind the tear-sheet, earnings-preview-beta, and funding-digest skills. Both are partner-built additions layered on top of the core modeling skills.
- Is the verify skill in this plugin used to check financial models?
- No, and the name makes that an easy mix-up. This repository's verify skill lives inside its Microsoft 365 install tooling and is for testing that tool's own admin scripts, such as clearing a stale Office add-in cache, against a sandboxed home directory. Checking a spreadsheet or financial model for errors is a job for audit-xls, the skill behind the debug-model command.
- Can I install just the equity research or private equity commands, without the rest?
- Yes. The vertical plugins are meant to be installed piecemeal: the financial-analysis core carries the shared modeling skills and the data connectors, and each additional vertical only adds the skills and commands specific to that workflow. See the listing page linked at the top of this guide for the exact plugin names to add.