Test

Uses Test persona. Take test execution output, parse it into structured results, produce a clear verdict, and recommend next actions.

Prerequisites: Test output provided (pasted or from execution)

Phase 1: Parse Output

Extract Structured Data

1. Identify the test framework from output format
2. Extract total, passed, failed, skipped counts
3. For failures, extract:
   • Test name
   • File and line location
   • Error message

Output

## Context Anchors

- **Phase:** 1 — Parse Output
- **Framework:** <detected framework>

## Execution Summary

| Metric  | Count |
| ------- | ----- |
| Total   | <N>   |
| Passed  | <N>   |
| Failed  | <N>   |
| Skipped | <N>   |

## Next Step

Continue to verdict.

**Approval Required:** No

Phase 2: Produce Verdict

Classify and Analyze

1. Apply verdict criteria:

2. For failures, analyze likely cause
3. Suggest remediation approach

Failure Analysis

For each failure:

1. Identify the failing test name
2. Extract the error message
3. Note the file and line if available
4. Analyze likely cause based on error type
5. Suggest remediation approach

Common Failure Patterns

Output

## Context Anchors

- **Issue:** #<number> (if applicable)
- **Test suite:** <name>

## Execution Summary

| Metric  | Count |
| ------- | ----- |
| Total   | <N>   |
| Passed  | <N>   |
| Failed  | <N>   |
| Skipped | <N>   |

## Verdict: <PASS | PARTIAL | FAIL>

<one paragraph summary>

## Failure Details (if any)

| Test   | Location    | Error     | Likely Cause |
| ------ | ----------- | --------- | ------------ |
| <name> | <file:line> | <message> | <analysis>   |

## Recommendations

- <next step 1>
- <next step 2>

## Next Step

<action based on verdict>

**Approval Required:** No

⛔ CHECKPOINT

STOP. Verify the verdict matches actual test output before reporting.

Critical: Never trust exit code alone. Parse and report actual counts.

Phase 3: TDD Workflow

Write Tests First

1. Read the ticket and implementation plan to understand what is being built
2. Write failing tests that define the expected behavior
3. Confirm tests fail for the right reason (not setup errors)
4. Hand off to Implementer — tests define the contract
5. Resume after implementation to verify tests pass

Phase 4: Coverage Analysis

Assess Test Coverage

1. Run the coverage tool specified in workspace.config.yaml
2. Identify untested paths — not just line coverage, but branch and error coverage
3. Flag high-risk untested areas (auth, validation, data transforms)
4. Report coverage delta if a baseline exists

Error Handling