Repository Analysis — `claude-for-financial-services`
claude-financial-services
Repository Analysis — claude-for-financial-services
This folder contains a structured deep-dive on the repo. Read in order:
| # | File | Topic |
|---|---|---|
| 0 | 00-overview.md | Purpose, problem, target users, what's in the repo |
| 1 | 01-tech-stack.md | Languages, frameworks, dependencies, runtime targets |
| 2 | 02-architecture.md | Components, data flow, key abstractions, ASCII/Mermaid diagrams |
| 3 | 03-entry-points.md | Where execution starts (Cowork install, Claude Code, Managed Agents API) |
| 4 | 04-design-patterns.md | "One source / two wrappers", task-decomposition isolation tiers, handoff protocol |
| 5 | 05-key-files.md | File-by-file responsibilities |
| 6 | 06-llm-usage.md | How LLMs are leveraged — prompts, guardrails, prompt-injection defenses |
| 7 | 07-other-considerations.md | Pitfalls, dual-use risks, things that surprised me |
Source references use path:line format throughout, e.g. scripts/orchestrate.py:40.