Generated: 2026-02-23
Total repos scanned: 91 (all repos returned by gh repo list)
Search method: Direct GitHub API content checks for each repo (not relying on GitHub code search, which returned zero results)
| File Type | Count | Repos |
|---|---|---|
| AGENTS.md | 13 | FINN-Web-App, Infrastructure, monitoring-infrastructure, github-infrastructure, greenhouse-pipeline-chat-worker, EWA-Actions, Statement-Fraud-Service-Serverless, experiment-service-sdk, roadmap-review-worker, Integrations, Terraform-Abstractions, greenhouse-chrome-extension |
| CLAUDE.md | 19 | Infrastructure, Tools, Tests, Account-Creation-Flow, github-infrastructure, google-workspace, Growthbook-Integration, EWA-Actions, Statement-Fraud-Service-Serverless, experiment-service-sdk, Finnwork, roadmap-review-worker, Integrations, Terraform-Abstractions, Manual-Paybacks-System, Payday-Prediction, banking-integrations, greenhouse-chrome-extension, financial-assistant-web-app |
| Both | 9 | Infrastructure, github-infrastructure, EWA-Actions, Statement-Fraud-Service-Serverless, experiment-service-sdk, roadmap-review-worker, Integrations, Terraform-Abstractions, greenhouse-chrome-extension |
| .cursorrules | 0 | — |
| copilot-instructions.md | 0 | — |
| Total unique repos with AI context | 22 out of 91 | ~24% coverage |
- Contains reusable Claude Code / Codex CLI skills (not per-repo AGENTS.md/CLAUDE.md)
- Skills:
setup-ralph-loop,cloudflare-worker,setup-release-please,finn-design-document-reviewer - README describes it as: "Repository for AI agent skills, AGENTS.md/CLAUDE.md files, and agent-related resources"
Role: Centralized CI/CD repo that syncs workflows/config across 70+ repos
AGENTS.md — Short, focused:
- Repo is centralized source of truth for CI/CD workflows and shared config
- Sync targets defined in
.github/sync-workflow-files.ymland.github/sync-other-files.yml - Edit templates here; downstream repos receive changes automatically
CLAUDE.md — Very comprehensive (200+ lines):
- Full architecture overview of file sync system
- Two-tier sync: workflow files + other config files
- Lists all workflow templates, custom actions, config templates
- Repository grouping by tech stack, team ownership, release requirements
- Coverage of 70+ repos
- Functions tests:
cd functions && npm test - Frontend tests:
ng test - Credolab SDK from private Cloudsmith registry (token handling notes)
AGENTS.md:
- Argo CD apps in
cd/k8s; sharedhelm-charts/ terraform/is legacy; uses Digger via PR comments- Repo effectively frozen for new infra
CLAUDE.md:
- ArgoCD deployment process details
- Helm chart/application structure
- GitOps-driven, no manual helm commands
- Internal tooling namespace patterns
- Root
.tffiles +modules/for reusable Terraform - Build:
make lint,pre-commit run -a,make docs - Digger-based CI/CD
- Conventional Commits required
AGENTS.md: Terraform modules for GitHub org management, Terratest (Go) suites, Digger workflow CLAUDE.md: Detailed module docs (finn-github-repository, finn-github-repository-ruleset, finn-github-team), CI/CD with Digger, development guidelines
- Cloudflare Worker for Greenhouse ATS
- Google Chat notifications + SLA monitoring
- Config in
config/pipeline.ymlcompiled tosrc/pipeline-config.ts
AGENTS.md: Lambda structure, test approach, Terraform workspaces, security
CLAUDE.md: Full architecture (5 Lambda functions), fraud detection rules, bank-specific validation, deployment flow
AGENTS.md: Short — keep behavior aligned across Node/TypeScript, Python, PHP SDKs
CLAUDE.md: Multi-language SDK architecture, development commands for all 3 languages, release management
AGENTS.md: Purpose, folder structure, credentials, scoring system, dev commands
CLAUDE.md: Extended version with two-stage flow, HTTP endpoints, reminder system
AGENTS.md: Comprehensive agent handbook — Nx monorepo, BPI/NTS/TVI service cheat sheets, coding standards
CLAUDE.md: Shorter version focused on quick-reference commands
AGENTS.md: Module organization, build/test commands, LocalStack testing
CLAUDE.md: Detailed module architecture (lambda, api-gateway, sqs, eventbridge, s3), design patterns, integration patterns
AGENTS.md: Build/test commands, worker/extension feedback loops
CLAUDE.md: 1Password secrets paths, URLs
- Collection of standalone utility tools
- Python linting with Ruff, JS funnel alerting
- GitHub Actions workflows for various automation
- Playwright test automation (354 lines)
- Multi-locale support (PH, TH, ZA, ID)
- Fixture-based testing, POM, database operations
- AWS Lambda functions for account creation
- Poetry + pytest, Terraform
- Detailed environment variables and secrets reference
- Terraform-managed Google Workspace
- Step-by-step user account provisioning
- Department codes, import procedures
- Lambda functions for experiment assignments
- MySQL + BigQuery dual-write
- LocalStack testing
- Application framework for microservices
- Kafka, Express.js, MySQL abstractions
- Service types: ApiService, MessageDrivenService
- Serverless payment processing (GBPrime, Xendit)
- Multi-country (TH, PH, ID)
- sandbox.sh for dev environments
- Laravel 9 payday prediction service
- PHPStan Level 9, comprehensive test suite
- Multi-country support
- Go + Gin + GORM banking HTTP service
- Clean architecture with provider pattern
- OpenAPI code generation
- AI financial assistant (OpenAI Agents SDK)
- Node.js + Express + Socket.io + PostgreSQL
- Heroku deployment
Automated-Charging, Statement-Service, netsuite-middleware, libraries, Serverless-Infrastructure, user-iam, Credit-Risk-Helpers, experiment-service, Serverless-Test, Serverless-Operations, agent-resources*, Risk-Evaluation-Temp, N8N-Workflows, line-integration, KTB-OCR-Service, Interview-Resources, Emails-Ingestion-Service, EKS-Configurations, Disbursement-Callbacks, Data-Science-Tools, B2C-Knack, B2B2C-Knack, Paybacks-Handling, experiment-service-sdk (already listed above — error), clevertap-aws-integration, Limit-Setting-Strategy, ELT-Data-Streams, cashtap-website, Advance-Mgt-System, SMS-Sending-Service, Core, platform-compute-infrastructure, Manual-Paybacks-System (already listed above), cloudflare-configuration, bigquery-aws-integration, Quickbooks-Connector, growthbook-examples, Payday-Prediction (already listed above), OCR-Service, Finnwork-Examples, analytical-base-table, Thai-ID-Card-OCR, Statement-Fraud-Service, Documentation, cost-explorer, strapi-cms, SMS-Payment-Links, FINN-Website, Console, synthetic-data-generator, Mission-Control, Local-Dev, Payday-Logics-Generation, Kasikornbank-ODD, Human-Resources, Accounting-Importer, Transfer-API, Risk-Evaluation-Backfilling, infrastructure-template, Decision-Engine, Experiments, Database-Access-Service, raycast-extensions, Advance-Fee-Calculator, github-runner, Charge-Times-Calculator, Branding, Anomaly-Detection, Landing-Page, Statement-Service-Old, Manual-Paybacks, Custom-Backend
*agent-resources has skills but no root AGENTS.md/CLAUDE.md
- 22 out of 91 repos (~24%) have AI context files — significant adoption
- CLAUDE.md is more common than AGENTS.md (19 vs 13) — Claude Code appears to be the primary AI coding tool
- 9 repos have both files — AGENTS.md tends to be shorter/focused, CLAUDE.md more comprehensive
- No .cursorrules or copilot-instructions.md found anywhere — Cursor and GitHub Copilot custom instructions not adopted
- EWA-Actions is the most comprehensive — serves as the org's CI/CD brain with full architecture docs
- agent-resources repo is a dedicated skills repository for Claude Code and Codex CLI
- Core, Console, Mission-Control — major production services — have NO AI context files yet
- GitHub code search returned zero results for all queries — unreliable for filename searches; direct API checks were necessary