/gate

/wicked-garden:crew:gate

Run quality engineering gate analysis on a target.

Arguments

• target (optional): File or directory to analyze. Default: current directory
• --gate (optional): Which gate to run. Default: strategy
  • value: Post-clarify - "Should we build this?"
  • strategy: Post-design - "Can we build it well?"
  • execution: Post-build - "Does it work?"
• --rigor (optional): Analysis depth. Default: standard
  • quick: Fast triage (~30s)
  • standard: Full analysis (~2min)

Instructions

1. Parse Arguments

Extract from the command arguments:

• target: The file or directory path (default: ".")
• gate: value, strategy, or execution (default: strategy)
• rigor: quick or standard (default: standard)

2. Validate Target

Check the target exists:

ls -la {target}

If target doesn't exist, inform user and ask for correct path.

3. Track Gate Task

Create a task for this gate analysis using TaskCreate:

TaskCreate(
  subject="QE: {project-name} - {gate} Gate on {target}",
  description="Running {gate} gate analysis on {target}",
  activeForm="Running {gate} gate analysis"
)

PreToolUse validates the task's metadata per scripts/_event_schema.py; the task persists natively under ${CLAUDE_CONFIG_DIR}/tasks/{session_id}/.

4. Invoke Orchestrator

Use Task tool to run qe-orchestrator with gate type:

Task(
  subagent_type="wicked-garden:crew:qe-orchestrator",
  prompt="""
  Run {gate} Gate analysis.

  Target: {target}
  Gate: {gate}
  Rigor: {rigor}

  Route to the appropriate gate orchestrator and return decision.
  """
)

5. Persist Gate Result

After the orchestrator returns, write the gate result to the project phase directory so that phase_manager.py can validate it during approval.

5a. Resolve the project directory:

sh "${CLAUDE_PLUGIN_ROOT}/scripts/_python.sh" "${CLAUDE_PLUGIN_ROOT}/scripts/_run.py" scripts/crew/phase_manager.py {project_name} current

This returns the current phase and project directory path.

5b. Write gate-result.json:

Write a JSON file to {project_dir}/phases/{current_phase}/gate-result.json with this schema:

{
  "result": "APPROVE|CONDITIONAL|REJECT",
  "gate": "{gate type}",
  "phase": "{current phase}",
  "reviewer": "wicked-garden:crew:qe-orchestrator",
  "score": 0.0-1.0,
  "findings": ["list of key findings"],
  "conditions": [{"id": "C-1", "description": "..."}],
  "timestamp": "ISO 8601"
}

• result: Extract from the orchestrator's decision (APPROVE, CONDITIONAL, or REJECT)
• score: Map the orchestrator's confidence/quality assessment to 0.0-1.0. Use 0.8 for clean APPROVE, 0.6 for CONDITIONAL, 0.3 for REJECT as defaults if no explicit score
• reviewer: Always "wicked-garden:crew:qe-orchestrator" (never use auto-approve names)
• conditions: Only include for CONDITIONAL results — list each condition with an id and description
• findings: Brief list of key observations from the gate analysis

6. Complete Gate Task

After writing the gate result, mark the gate task as completed:

TaskUpdate(taskId={task_id from Step 3}, status="completed")

7. Format Output

Display the result from orchestrator:

## QE Analysis Complete

**Target**: {target}
**Gate**: {gate}
**Rigor**: {rigor}

---

{orchestrator result}

---

**Task**: {task_id} (if created)

To view full evidence: `/wicked-garden:crew:evidence`

8. Handle Decisions

Based on the decision:

APPROVE:

Gate: APPROVE

Ready to proceed. No blocking issues found.

CONDITIONAL:

Gate: CONDITIONAL

Conditions to address:
1. {condition 1}
2. {condition 2}

You can proceed, but these should be addressed.

REJECT:

Gate: REJECT

Blockers found:
1. {blocker 1}

Must fix before proceeding.

9. Show Evidence Summary

After displaying the gate result, show attached evidence:

---

### Evidence Attached

| Type | Artifact |
|------|----------|
| Gate Result | `L3:qe:{gate}-gate` |
| Decision Memory | stored in wicked-garden:mem (if available) |

To view full evidence: `/wicked-garden:crew:evidence {task_id}`

Examples

# Default: Strategy gate on current directory
/wicked-garden:crew:gate

# Value gate on outcome document
/wicked-garden:crew:gate outcome.md --gate value

# Strategy gate on specific directory
/wicked-garden:crew:gate src/auth --gate strategy

# Execution gate after implementation
/wicked-garden:crew:gate src/auth --gate execution

# Quick triage
/wicked-garden:crew:gate src/payments --rigor quick

Gate Selection Guide

Phase Just Completed	Gate to Run
clarify	--gate value
design	--gate strategy
build	--gate execution