Peak POS internal documentation
Use this site for repository operations, architecture, support procedures, API ownership, security, and validation. Source code, Bazel targets, checked-in scripts, and GitHub workflows remain the implementation source of truth.
Develop and validate
Start with the page that matches the task:
- Local development and manual QA — start the local stack, find ports, run smoke paths, rebuild a component, or connect a simulator.
- Testing strategy — choose focused Bazel targets, understand remote execution, and record evidence.
- Repository operations — build targets, CI workflows, environment boundaries, and operator-only actions.
- Database reset and reinstall — reset local validation state or perform an explicitly authorized Cloud Spanner rebuild.
- Pre-commit hooks — install and run
prekas a supplement to Bazel validation. - Peak POS SDK release checklist — verify SDK artifacts and consumers before using a publishing workflow.
The supported local command surfaces are scripts/local-validation.sh for stack lifecycle and scripts/test-local.sh for automated batteries. Retired QA ledgers and obligation-rollup commands are not release gates.
Support operations
API contracts
Generated OpenAPI artifacts and executable specifications take precedence over copied request or response examples.
Architecture
- Peak re-architecture and owner map
- Android payment interface
- Browser scanner inventory research
- Browser scanner inventory implementation plan
- Gateway SDK capability matrix
- Demo provisioning
- Account and tax gateway alignment
- Azure staging status
- Azure production deployment plan
- Terrakube GCP ownership
Use the owner map before changing a route, data owner, payment boundary, or deployment lane. Historical plans describe intent, not necessarily the current runtime.
Security and compliance
- Security audit
- Security remediation
- Penetration test report
- POS audit coverage
- Age policy matrix
- PCI redaction audit
QA and end-to-end coverage
- QA checklist index
- Command-surface verification
- Documentation build and render verification
- Evidence summary
- Hardware smoke runbook
- Playwright MCP testing
- Playwright behavior coverage gaps
A report or checklist records evidence; it does not replace the command that produced that evidence. Keep local, staging, hardware, and production results labeled separately.