analyzing-codebaselisted

# Analyzing the codebase Build a technical map of the repository grounded in evidence: every claim traces to a file, dependency, or config entry actually found. See [references/stack-detection.md](../../references/stack-detection.md) for the detection patterns this skill applies across frontend, backend, database, testing, and infrastructure, and [enforcing-safety-baseline](../enforcing-safety-baseline/SKILL.md) for the evidence and secret-hygiene rules this skill follows — a config file scanned during discovery may contain a real credential; it gets flagged by location, never reproduced in full. ## Workflow 1. **Survey structure fast.** `Glob` for manifest/config files (package.json, composer.json, requirements.txt, angular.json, playwright.config.*, docker-compose.yml, etc.) before reading full source trees. Determine monorepo vs. single-app from workspace/package manager signals. 2. **Detect the stack.** Apply `stack-detection.md` patterns for frontend, backend, database/persistence, testing, and infrastructure. Report only technologies with actual evidence; when evidence is ambiguous, read the specific config to disambiguate rather than guessing. 3. **Map architecture.** Identify layering (e.g. controllers/services/repositories, feature-based modules, monorepo packages), how frontend and backend communicate (REST/GraphQL/RPC contracts), and where cross-cutting concerns live (auth, validation, error handling). 4. **Identify persistence and tenancy.** Locate schema/migra