/validate - Implementation Validation

Verify that implementation meets acceptance criteria and passes tests.

Purpose

Ensure quality by:

• Running test suites
• Checking acceptance criteria
• Validating artifact schemas
• Reporting findings

Inputs

• $ARGUMENTS: Task ID or artifact path to validate
• BACKLOG for acceptance criteria: docs/development/BACKLOG.md
• PRD for requirements: docs/architecture/PRD.md
• ${PROJECT_NAME}: Current project context

Outputs

Validation report stored in Serena memory (transient).

Workflow

1. Identify Target

Parse $ARGUMENTS:

• Task ID → Load task details and AC from BACKLOG
• File path → Validate specific artifact
• No args → Validate most recent implementation

2. Load Acceptance Criteria

From task details or PRD:

• List all criteria to check
• Note any dependencies

3. Run Tests

Execute appropriate test commands:

# Common patterns
npm test
pytest
make test
go test ./...

Capture results: passed, failed, skipped.

4. Verify Acceptance Criteria

For each criterion:

• Can it be verified automatically?
• If yes, run verification
• If no, check manually via code inspection

5. Check Quality

Run quality checks if available:

# Linting
npm run lint
ruff check .

# Type checking
tsc --noEmit
mypy .

# Format check
prettier --check .
ruff format --check .

6. Validate Artifacts

If artifacts were produced:

• Check schema compliance
• Verify required fields
• Validate references

7. Report Findings

Create validation report:

---
date: YYYY-MM-DD
task: [Task ID]
target: [What was validated]
status: pass | fail | partial
---

## Summary
[Overall status and key findings]

## Test Results
| Suite | Passed | Failed | Skipped |
|-------|--------|--------|---------|
| [Name] | X | Y | Z |

## Acceptance Criteria
| ID | Criterion | Status | Evidence |
|----|-----------|--------|----------|
| AC1 | [Description] | Pass/Fail | [How verified] |

## Quality Checks
- [ ] Linting: [status]
- [ ] Type check: [status]
- [ ] Format: [status]

## Issues Found
1. [Issue description]
   - Severity: Low/Medium/High
   - Location: [file:line]

## Recommendation
[Pass/Fix issues/Block deployment]

Validation Levels

Quick Validation

• Run tests
• Check critical AC only
• Fast feedback

Full Validation

• All tests
• All AC
• Quality checks
• Artifact validation

Pre-Deploy Validation

• Full validation
• Security checks
• Performance baseline
• Integration tests

Policy References

Should-read from ~/.claude/policy/RULES.md:

• Failure Investigation - Root cause analysis, never skip tests or validation