Second Line Financial Services

Why

Anthropic launched financial services plugins, but the initial set is all geared toward the front line. Pitchbooks, valuations, fund admin, finance-ops.

What the sector is still missing Claude skills for is the governance, risk, compliance, and controls side. The credible challenge layer.

What

Plugins for second-line work in financial services.

• Risk and compliance core (scoping, obligation mapping, control matrices, issue write-ups, policy-gap reviews)
• Regulatory change management (rule extraction, impact assessments, exam management)
• AI governance and model risk (use-case intake, risk tiering, validation plans, agentic and GenAI controls)
• Third party and operational resilience (vendor diligence, contract review, resilience testing)
• Compliance testing (test plans, sampling, workpapers, QA)
• Risk reporting (committee packs, KRI commentary)
• Consumer compliance and fair lending (adverse actions, UDAAP, complaints)

And then there are the sector layers, the specificities that differ across banking, insurance, capital markets and asset management, payments and fintech. Not to mention cross-cutting skills like cyber, privacy, conduct, AI ethics, and climate.

Where it stands

Live on GitHub. Add them as Agent Skills in Claude, or Codex, or anything else that reads SKILL.md. Each skill cites the actual regs and standards it pulls from, so Claude isn’t guessing, but I wouldn’t trust them to run full workflows yet. They’re not validated, and nothing’s governing them. You could point the AI governance skill at itself, and at the rest of these skills… fun recursion, but no, haven’t done that yet.