You are an AI coding agent working in this repository. Your job is not “write code”; it is ship reviewable, testable changes with evidence. Every pull request must make the implementation and user experience unambiguous.
- Behavior-first summary Provide a short “What changed” list focused on user-visible behavior and system contracts (UI, API, data, permissions).
- User Flow Coverage
For every feature or change, document:
- Happy path
- Failure paths (validation errors, permissions, network/offline, empty states, 404/500, timeouts)
- Biggest pain point are email notifications. They are flaky and hard to test and always have bugs.
- Update /test page for easy manual tests.
- Evidence
Include at least one of:
• Screenshot(s) for UI changes
• Short screen recording (10–45s) showing the flow
• CLI output snippet for non-UI changes (commands + results) - Test Proof
Tests must exist and must be runnable locally/CI. Provide:
• What tests you added/updated
• The exact commands to run
• The result summary (pass/fail, coverage deltas if available) - Act as an architect and review the PR for areas of weakness. Identify those clearly
- Act as a tester and identify all of the edge cases. Write tests. Prefer integration tests that make DB and API calls.
- Code coverage must be measured and maintained.
- Must have thorough monitoring
- Must include thorough logging
- Was canonical Agents.md followed? Must be true.
- Was code coverage report included? Must be true.