shipkit-test-cases

shipkit-test-cases - Code-Anchored Test Case Management

Purpose: Generate and maintain test case specifications anchored to source files, with automatic staleness detection when code changes. Optimized for AI execution during team implementation.

---

When to Invoke

User triggers:

• "Generate test cases"
• "What should we test?"
• "Set up test coverage"
• "Create test spec"
• "Test cases for [feature]"

Workflow position:

• After specs exist (helpful but not required)
• Before team implementation (provides execution targets)
• Can run standalone against existing codebase

---

Prerequisites

Soft requirements (helpful but not blocking):

• .shipkit/specs/ — Helps derive cases from requirements
• .shipkit/product-discovery.json — Helps infer priority from user personas

No hard requirements — Can analyze code directly without specs

---

Process

Completion Tracking (MANDATORY)

After determining scope (Step 1) and identifying feature areas, create tasks:

1. For EACH feature area identified: TaskCreate: "Generate cases: {feature-name}"
2. TaskCreate: "Write STRATEGY.md"
3. For EACH feature: TaskCreate: "Write cases/{feature}.md"
4. TaskCreate: "Write coverage.json"
5. TaskCreate: "Enrich UI test cases with selectors (if applicable)"
6. TaskCreate: "Archive existing artifact (if exists)"
7. TaskCreate: "Output summary with metrics"

Rules:

• Do NOT dispatch test case generation once for all features — generate per feature area
• TaskUpdate each task to completed only after the file is written to disk
• STRATEGY.md, cases/*.md, AND coverage.json must ALL be written — any one alone is not "done"
• Do NOT present the summary until all 3 output types are confirmed written

Step 1: Determine Scope

If user specifies feature/scope:

• Focus on that feature's source files
• Read relevant spec if exists

If no scope specified:

• Scan src/ for code structure
• Check git diff for recently changed files
• Prioritize uncovered code