# Create Gate
## System Message
You are a quality gate architect specializing in gate design for the Claude Prompts MCP system. Your task is to create production-quality validation gates that match the standards of existing gates like code-quality and educational-clarity.
## User Message
{% if tool_gate_builder %}
{# Auto-execute workflow: gate input was provided and validated #}
{% if tool_gate_builder.valid %}
{% if tool_gate_builder_result %}
{# Auto-execute completed successfully #}
## Gate Created Successfully
{{ tool_gate_builder_result.text }}
**Summary:**
- **ID:** `{{ tool_gate_builder.auto_execute.params.id }}`
- **Name:** {{ tool_gate_builder.auto_execute.params.name }}
- **Type:** {{ tool_gate_builder.auto_execute.params.type }}
- **Pass Criteria:** {{ tool_gate_builder.summary.pass_criteria_count }}
- **Severity:** {{ tool_gate_builder.summary.severity }}
{% if tool_gate_builder.warnings | length > 0 %}
### Warnings (non-blocking)
{% for warning in tool_gate_builder.warnings %}
- {{ warning }}
{% endfor %}
{% endif %}
**Next Steps:**
1. Test the gate: `prompt_engine(">>your_prompt :: '{{ tool_gate_builder.auto_execute.params.id }}'")`
2. List gates: `system_control(action:"gates", operation:"list")`
{% else %}
{# Validation passed but auto-execute not available #}
## Gate Validated
The gate definition passed validation. Call `resource_manager` with the following parameters:
```json
{{ tool_gate_builder.auto_execute.params | dump(2) }}
```
{% if tool_gate_builder.warnings | length > 0 %}
### Warnings (non-blocking)
{% for warning in tool_gate_builder.warnings %}
- {{ warning }}
{% endfor %}
{% endif %}
{% endif %}
{% else %}
{# Validation failed #}
## Validation Failed
The gate definition has errors that need to be fixed:
{% for error in tool_gate_builder.errors %}
- {{ error }}
{% endfor %}
{% if tool_gate_builder.warnings | length > 0 %}
### Warnings
{% for warning in tool_gate_builder.warnings %}
- {{ warning }}
{% endfor %}
{% endif %}
Please fix the errors above and try again with corrected gate definition.
{% endif %}
{% else %}
{# Design workflow: no gate input provided, guide LLM to design #}
Create a new quality gate with the following specifications:
**Name:** {{name}}
**Purpose:** {{purpose}}
{% if gate_type %}**Type:** {{gate_type}}{% endif %}
---
## Reference: Existing Gate Examples
### Example 1: Code Quality Gate
**gate.yaml:**
```yaml
id: code-quality
name: Code Quality Standards
type: validation
description: Ensures generated code follows best practices and quality standards
guidanceFile: guidance.md
pass_criteria:
- type: content_check
min_length: 100
required_patterns:
- try
- catch
- //
- function
- const
forbidden_patterns:
- TODO
- FIXME
- hack
- password123
- type: pattern_check
regex_patterns:
- 'function\s+\w+\s*\('
- '\/\/.*\w+'
- 'try\s*\{'
retry_config:
max_attempts: 2
improvement_hints: true
preserve_context: true
activation:
prompt_categories:
- code
- development
explicit_request: false
```
**guidance.md:**
```markdown
**Code Quality Standards:**
- Include error handling and input validation
- Add inline comments for complex logic
- Follow consistent naming conventions
- Consider edge cases and boundary conditions
- Optimize for readability over cleverness
- Include basic documentation/docstrings
- Follow security best practices
```
### Example 2: Educational Clarity Gate
**gate.yaml:**
```yaml
id: educational-clarity
name: Educational Content Clarity
type: validation
description: Ensures educational content is clear, well-structured, and pedagogically sound
guidanceFile: guidance.md
pass_criteria:
- type: content_check
min_length: 400
required_patterns:
- example
- step
- understand
- learn
- type: pattern_check
keyword_count:
example: 2
for example: 1
regex_patterns:
- '^##\s+.*$'
- '^\d+\.\s+.*$'
- '^\s*[-*]\s+.*$'
retry_config:
max_attempts: 2
improvement_hints: true
preserve_context: true
activation:
prompt_categories:
- education
- documentation
explicit_request: false
```
---
## Schema Reference
### gate.yaml Fields
| Field | Type | Required | Description |
| ----------------- | ------ | -------- | ----------------------------------------------------- |
| `id` | string | Yes | Lowercase-hyphenated identifier |
| `name` | string | Yes | Human-readable name |
| `type` | enum | Yes | `validation` (pass/fail) or `guidance` (advisory) |
| `description` | string | Yes | Brief description of what the gate validates |
| `guidanceFile` | string | No | Path to guidance.md (default: 'guidance.md') |
| `guidance` | string | No | Inline guidance content (alternative to file) |
| `severity` | enum | No | `critical`, `high`, `medium`, `low` (default: medium) |
| `enforcementMode` | enum | No | `blocking`, `advisory`, `informational` |
| `pass_criteria` | array | No | Validation criteria definitions |
| `retry_config` | object | No | Retry behavior configuration |
| `activation` | object | No | When to activate the gate |
### Pass Criteria Types
| Type | Fields | Description |
| ------------------------ | ------------------------------------------------------------- | ------------------------------------ |
| `content_check` | min_length, max_length, required_patterns, forbidden_patterns | Check content length and patterns |
| `pattern_check` | regex_patterns, keyword_count | Regex matching and keyword frequency |
| `llm_self_check` | prompt_template, pass_threshold | LLM-based validation |
| `methodology_compliance` | methodology, min_compliance_score, quality_indicators | Framework compliance |
### Retry Configuration
```yaml
retry_config:
max_attempts: 2 # How many retry attempts allowed
improvement_hints: true # Include hints for improvement
preserve_context: true # Keep context between retries
```
### Activation Rules
```yaml
activation:
prompt_categories: # Categories that trigger this gate
- code
- development
frameworks: # Frameworks that trigger this gate
- cageerf
explicit_request: false # Only activate when explicitly requested
```
---
## Your Task
Based on **{{name}}** with purpose "{{purpose}}":
1. **Choose gate type**: {% if gate_type %}{{gate_type}}{% else %}validation or guidance{% endif %}
2. **Write clear description** explaining what the gate validates
3. **Define pass_criteria** with appropriate types for your use case
4. **Create guidance content** with actionable standards
5. **Configure activation** for appropriate prompt categories
---
## How to Create the Gate
Make a **second `prompt_engine` call** to `>>create_gate` with all fields in the `options` parameter. The script tool will automatically detect the schema match and create the gate files.
**Example call:**
```text
prompt_engine(
command: ">>create_gate",
options: {
"id": "your-gate-id",
"name": "Your Gate Name",
"type": "validation",
"description": "What this gate validates",
"guidance": "**Standards:**\n- Criterion 1\n- Criterion 2",
"pass_criteria": [...],
"activation": {...}
}
)
```
---
## Required Fields (in `options`)
| Field | Type | Description |
| ------------- | ------ | ------------------------------------------------------------------- |
| `id` | string | Lowercase-hyphenated identifier (e.g., "api-docs", "test-coverage") |
| `name` | string | Human-readable name |
| `type` | enum | `validation` or `guidance` |
| `description` | string | What the gate validates |
## Optional Fields (for full configuration)
| Field | Type | Description |
| ----------------- | ------ | ----------------------------------------------------- |
| `guidance` | string | Inline guidance content (markdown) |
| `guidanceFile` | string | Path to guidance.md file |
| `severity` | enum | `critical`, `high`, `medium`, `low` |
| `enforcementMode` | enum | `blocking`, `advisory`, `informational` |
| `pass_criteria` | array | Validation criteria objects |
| `retry_config` | object | `{max_attempts, improvement_hints, preserve_context}` |
| `activation` | object | `{prompt_categories, frameworks, explicit_request}` |
---
## JSON Structure Reference
```json
{
"id": "<lowercase-hyphenated>",
"name": "<Human Readable Name>",
"type": "validation",
"description": "What this gate validates",
"severity": "medium",
"enforcementMode": "blocking",
"guidance": "**Standards:**\n- Criterion 1\n- Criterion 2\n- Criterion 3",
"pass_criteria": [
{
"type": "content_check",
"min_length": 100,
"required_patterns": ["pattern1", "pattern2"],
"forbidden_patterns": ["bad_pattern"]
},
{
"type": "pattern_check",
"regex_patterns": ["regex1", "regex2"],
"keyword_count": {
"keyword": 2
}
}
],
"retry_config": {
"max_attempts": 2,
"improvement_hints": true,
"preserve_context": true
},
"activation": {
"prompt_categories": ["category1", "category2"],
"frameworks": [],
"explicit_request": false
}
}
```
**Requirements:**
- Gate IDs: lowercase-hyphenated (e.g., `api-docs`, `test-coverage`)
- Type: Must be `validation` or `guidance`
- Either `guidance` (inline) or `guidanceFile` (path to .md file)
- pass_criteria: At least one criterion for validation gates
- Regex patterns: Must be valid regular expressions
{% endif %}
