UAT Testing

Purpose

This skill performs User Acceptance Testing (UAT) to validate that a feature or system behaves correctly from the end-user’s perspective.

It focuses on:

• Realistic user workflows
• End-to-end behavior
• Edge cases and failure handling
• Usability and UX friction
• System correctness under real usage

This is not unit testing. This is not architectural review. This is behavioral validation at the system boundary.

This skill must honor governance rules defined in .agents/profile.yaml if present. If the active profile enables automatic issue creation from findings (e.g., auto_create_issue_from_assessment or equivalent flag), any UAT defect at or above the configured severity threshold must trigger issue creation via the track-issues skill. This skill must not create issue files directly.

When To Use

Use this skill:

• After execution-audit has been performed and implementation is verified.
• Before major release milestones

UAT Workflow

1. Identify target persona(s).

2. Define primary user flows.

3. Define edge-case scenarios.

4. Execute flows (manually or via commands where applicable).

5. Record expected vs actual behavior.

6. Identify UX friction or unclear messaging.

7. Determine PASS / CONDITIONAL PASS / FAIL.

8. Recommend corrective actions if needed.

9. Apply profile severity policy to identified defects.

10. If auto issue creation is enabled by profile, invoke the track-issues skill to create or update corresponding pending issues for qualifying defects.

11. If previously created pending issues are validated as resolved during UAT, invoke the track-issues skill to transition those issues from pending/ to resolved/.

Test Design Principles

• Test from the outside-in.
• Simulate real usage, not idealized usage.
• Include at least one failure scenario per major flow.
• Explicitly define expected results before execution.
• Record observable behavior only.
• Avoid speculative judgments without evidence.

UAT Report Template

# UAT Report — <Feature or Topic>

## Persona
<Who is the user being simulated?>

## Environment
- Version:
- Branch/Commit:
- Platform:

---

## Primary Scenarios

### Scenario 1: <User Goal>

Steps:
1.
2.
3.

Expected:
<expected outcome>

Actual:
<observed behavior>

Result: PASS / FAIL

---

### Scenario 2: <User Goal>
...

---

## Edge Cases

- Case:
  - Expected:
  - Actual:
  - Result:

---

## Error Handling Evaluation

- Input validation clarity
- Error message usefulness
- Recovery guidance

---

## UX Observations

- Confusing flows
- Friction points
- Missing feedback
- Overly complex steps

---

## Defects Identified

- Defect:
  - Title:
  - Severity: Low / Medium / High
  - Production Blocking: Yes / No
  - Reproduction Steps:
  - Impact:
  - Suggested Fix:

---

## Final UAT Status
PASS / CONDITIONAL PASS / FAIL

Recommendation:
- <Move issue to resolved?>
- <Required fixes before release?>

Completion Standard

UAT is complete when:

• At least one primary user flow is validated.

• At least one edge case is tested.

• Expected vs actual behavior is documented.

• PASS/FAIL decision is explicit.

• If profile enables auto issue creation, all qualifying defects must be delegated to the track-issues skill.

• If UAT confirms that related pending issues are resolved, those issues must be transitioned to resolved/ via the track-issues skill.

repo_root/agentspace/uat/YYYY-MM-DD-<topic>.md

Rules:

• Use ISO date format (YYYY-MM-DD).
• Do not overwrite existing UAT reports unless explicitly instructed.
• UAT must precede moving an issue to resolved.