openspec-test

# OpenSpec Test

Layered verification engine for OpenSpec changes. Runs tests in order: smoke → integration → manual → PBT.

## Workflow: Detect → Run Layers → Report

```mermaid
flowchart LR
    A["Detect framework"] --> B["Smoke tests"]
    B --> C{"Pass?"}
    C -->|Yes| D["Integration tests"]
    C -->|No| X["Report failure"]
    D --> E{"Pass?"}
    E -->|Yes| F["Manual instructions"]
    E -->|No| X
    F --> G["PBT (if applicable)"]
    G --> H["Report summary"]
    X --> H

    classDef action fill:#C8E6C9,stroke:#388E3C,color:#000
    classDef decision fill:#FFF9C4,stroke:#FBC02D,color:#000
    classDef fail fill:#FFCDD2,stroke:#D32F2F,color:#000
    class A,B,D,F,G,H action
    class C,E decision
    class X fail
```

**Critical**: Stop on layer failure. No point running integration if smoke fails.

## Exploration Strategy

Before testing, consult `openspec/project.md` → Exploration Strategy section:

1. **Context sources**: Read `primary` files (project.md, proposal.md, specs)
2. **Must-read files**: CLAUDE.md, settings.json (project constraints)
3. **Test config**: Detect test framework from package.json, pytest.ini, cargo.toml, etc.
4. **Existing tests**: Grep for test files to understand project test patterns
5. **Philosophy**: Read Execution Philosophy section for current mode

## Test Layers

```yaml
layers:
  smoke:
    purpose: Basic health checks
    examples: ["app starts", "endpoints respond", "no crash on basic input"]
    auto: true

  integration:
    purpose: Component interactions
    examples: ["API contracts match", "DB state consistent", "error handling"]
    auto: true

  manual:
    purpose: Human-verified critical paths
    examples: ["UI flows", "user journeys", "edge cases needing judgment"]
    auto: false  # Generate instructions only

  pbt:
    purpose: Property-based edge case discovery
    examples: ["invariants hold", "no unexpected states", "boundary conditions"]
    auto: true
    frameworks: [hypothesis, fast-check, propte