version: "1.0.0" name: test-driven description: TDD workflow for go-adk-q. Write the failing test first, implement the minimum code to pass, then refactor. Covers unit, integration, and table-driven patterns used in this codebase. compatibility: Designed for go-adk-q — a Google ADK Go reference implementation with Bubbletea TUI.

Test-Driven Development — go-adk-q

Use this skill when adding a new feature, fixing a bug, or refactoring. Write the test first. Always.

The cycle

RED   → write a failing test that describes the desired behaviour
GREEN → write the minimum code to make it pass
CLEAN → refactor without breaking the test

Never write implementation code without a failing test first. If you catch yourself thinking "I'll add the test after" — stop and reverse.

Project test conventions

• Test files live next to the code they test: foo.go → foo_test.go
• Use package main (white-box) for TUI internals; package foo_test (black-box) for public APIs
• Table-driven tests with t.Run:

``go tests := []struct { name string input string want string }{ {"empty", "", ""}, {"plain text", "hello", "hello\n"}, } for _, tt := range tests { t.Run(tt.name, func(t *testing.T) { got := renderMarkdown(s, tt.input, 80, lipgloss.NewStyle()) if got != tt.want { t.Errorf("got %q, want %q", got, tt.want) } }) } ``

• Use t.Helper() in shared assertion helpers
• Use -race flag when testing anything concurrent: go test -race ./...

Adding a new TUI feature — step by step

1. Describe the behaviour in plain English before touching code.

Example: "When the user types /skills, the slash menu should show it as an option."

1. Write the test in cmd/tui/:

``go func TestSlashMatchesSkills(t *testing.T) { matches := slashMatches("/sk") found := false for _, m := range matches { if m.name == "/skills" { found = true } } if !found { t.Error("expected /skills in slash matches for prefix /sk") } } ``

1. Run it — watch it fail:

`` go test ./cmd/tui/ -run TestSlashMatchesSkills -v `` A test that never fails is not a test.

1. Implement the minimum code to make it pass.

1. Run the full suite — no regressions allowed:

`` go test ./cmd/tui/ -count=1 ``

1. Refactor if needed, keeping tests green.

Bug fix workflow

1. Reproduce first. Write a test that fails because of the bug.
2. Commit the failing test (optional but recommended — proves the bug existed).
3. Fix the bug. The test goes green.
4. Run full suite. Ship.

Never fix a bug without a regression test. The next person to touch that code will thank you.

What to test in this codebase

What NOT to unit-test

• The full tea.Program lifecycle (too much setup, no value)
• HTTP calls to real providers (use integration tests or manual smoke tests)
• copyToClipboard (system call — test with the QA skill instead)

Running tests

go test ./...                    # all packages
go test ./cmd/tui/ -v            # TUI tests with names
go test ./cmd/tui/ -run TestMd   # filter by name prefix
go test -race ./...              # race detector
go test -cover ./...             # coverage report

Coverage goal: keep cmd/tui/ above 60%. Check with:

go test -coverprofile=cov.out ./cmd/tui/ && go tool cover -func=cov.out