Test

The Test persona drives this workflow. 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 from raw test output.

Steps

1. Identify the test framework from output format
2. Extract total, passed, failed, skipped counts
3. For failures, extract test name, location, error message

Phase 2: Produce Verdict

Classify the result and provide analysis.

Verdict Criteria

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
5. Suggest remediation approach

Common Failure Patterns

Output

## Context Anchors

- **Issue:** #<number> - <title> (if applicable)
- **Suite:** <test suite name>

## Test Results

| Suite | Total | Pass | Fail | Skip |
| ----- | ----- | ---- | ---- | ---- |
| ...   | ...   | ...  | ...  | ...  |

## Verdict: PASS | PARTIAL | FAIL

### Failures (if any)

| Test   | Error           | Location    | Likely Cause   |
| ------ | --------------- | ----------- | -------------- |
| <name> | <error message> | <file:line> | <likely cause> |

### Recommendations

- <next action>

## Next Step

<what comes next>

**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.

Error Handling