feat: intelligent test failure diagnosis and categorization

[thomhurst]

Description

When tests fail, error messages are basic with no intelligent diagnosis. There's no automated categorization or suggested fixes.

Why

Developers spend time debugging obvious failures. There's no way to automatically distinguish between test failures and infrastructure problems.

Proposed Approach

• Create ITestFailureDiagnostics analyzing exception type, test dependencies, resource state
• Implement TestFailureCategorizer: InfrastructureFail, TestCodeFail, ProductionFail, EnvironmentFail
• Generate diagnostic suggestions, e.g.: "This test timed out after 5s. The slow operation was Database.Query(). Consider [Timeout(10000)]."
• Show dependency chain when a test fails due to a dependency failure

[claude]

Issue Analysis

Type: Feature request

Priority: Medium — improves developer experience and debuggability, but no existing functionality is broken.

Related issue (not a duplicate): This overlaps in scope with #4888 (feat: test timeout diagnostics with stack traces and deadlock detection), which covers a narrow subset of what's proposed here (timeout failures specifically). The categorization system in this issue — InfrastructureFail, EnvironmentFail, etc. — could naturally incorporate timeout diagnostics from #4888 as one category. Worth considering whether these should be designed together or if #4888 becomes a building block for this broader system.

Scope notes: This is a significant feature requiring new abstractions (ITestFailureDiagnostics, TestFailureCategorizer), a suggestion engine, and dependency chain reporting. Not a simple first contribution — coordinating with the timeout diagnostics work in #4888 would be advisable before starting.