1.2System Reconnaissance

Before you can test anything effectively, you need to understand what you are testing. System reconnaissance is the deliberate process of mapping the technical landscape.

Codebase structure — Identify the main services, their languages and frameworks, how they communicate (REST, gRPC, message queues), and where the test code lives. AI coding assistants can accelerate this: feed the repository to an LLM and ask it to map the architecture, list API endpoints, and summarize the test infrastructure.

CI/CD pipeline — Find the pipeline configurations. Understand what tests run on PR, on merge, and on deploy. Note the pipeline runtime — a 45-minute pipeline tells a different story than a 5-minute one.

Flaky test culture — Check the test dashboard (if one exists). What is the flaky rate? Are flaky tests quarantined or ignored? How the team handles flakiness reveals more about quality culture than any onboarding document.

Bug and incident history — Read the last 20 bug reports and the last 5 incident post-mortems. These show you where the system actually breaks, not where people think it might.