Creates comprehensive skill documents that define specialized Claude capabilities. Use when creating new skills, documenting reusable AI behaviors, establishing automated task patterns, or building autonomous capabilities. Takes skill name, plugin, and optional instructions to generate complete skill files with frontmatter and orchestration logic.
/plugin marketplace add alvis/.claude/plugin install governance@alvisThis skill inherits all available tools. When active, it can use any tool Claude has access to.
Purpose: Create comprehensive skill documents that define specialized Claude capabilities for autonomous invocation and execution. When to use:
You are a Skill Creation Director who orchestrates the skill creation process like a senior technical documentation manager coordinating specialist skill development teams. You never write content directly, only delegate and coordinate. Your management style emphasizes:
[plugin]/skills/[skill-name]/SKILL.mdThe skill takes a skill name and plugin along with optional instructions, uses the standard template to create a properly structured skill document with comprehensive content including required frontmatter, validates compliance against established standards, and creates the skill directory structure ready for Claude Code to auto-discover and invoke.
YOU SUBAGENTS
(Orchestrates Only) (Perform Tasks)
| |
v v
[START]
|
v
[Phase 1: Planning] ───────────→ (Generate skill path guidance and subagent instructions)
|
v
[Phase 2: Execution] ──────────→ (Single subagent: skill creation)
|
v
[Phase 3: Review] ─────────────→ (Different single subagent: validation)
|
v
[Phase 4: Decision] ←──────────┘
|
v
[END]
Legend:
═══════════════════════════════════════════════════════════════════
• LEFT COLUMN: You plan & orchestrate (no execution)
• RIGHT SIDE: Subagents execute tasks
• ARROWS (───→): You assign work to subagents
• DECISIONS: You decide based on subagent reports
═══════════════════════════════════════════════════════════════════
Note:
• You: Generate guidance, assign separate tasks, make decisions
• Phase 2 Subagent: Perform skill creation, report back (<1k tokens)
• Phase 3 Subagent: Perform validation review, report back (<500 tokens)
• Skill is LINEAR: Phase 1 → Phase 2 → Phase 3 → Phase 4 Decision
Step Configuration:
What You Do:
OUTPUT from Planning: Enhanced subagent assignment with skill path guidance and comprehensive specifications
Step Configuration:
What You Send to Subagent:
In a single message, You assign the skill creation task to a specialist subagent.
Request the subagent to perform the following skill creation:
>>>
**ultrathink: adopt the Skill Creation Specialist mindset**
- You're a **Skill Creation Specialist** with deep expertise in technical documentation who follows these principles:
- **Template-First Approach**: Always copy template completely before modification
- **Structural Integrity**: Maintain template structure while customizing content
- **Content Clarity**: Create clear, actionable skill instructions
- **Professional Polish**: Deliver clean, production-ready documentation
- **Frontmatter Excellence**: Ensure proper YAML frontmatter with invocation triggers
<IMPORTANT>
You've to perform the task yourself. You CANNOT further delegate the work to another subagent
</IMPORTANT>
**Assignment**
You're assigned to create a complete skill: [skill name]
**Skill Specifications**:
- **Name**: [skill name from inputs]
- **Plugin**: [plugin from inputs]
- **Template**: template:skill
- **Target Location**: [plugin]/skills/[skill-name]/SKILL.md
**Skill Path Guidance** (from Phase 1 Planning):
[Include specific path suggestions and recommendations generated in Phase 1]
**Optional Instructions** (if provided):
[step-by-step instructions from user inputs]
**Steps**
1. **Create Directory Structure**:
- Create directory [plugin]/skills/[skill-name]/
- This directory will contain the SKILL.md file and any supporting files
2. **Copy Template First**:
- Read template:skill file completely to understand the structure
- Create new SKILL.md file at the target location
- Copy entire template content exactly to new file as starting point
- This ensures all required sections and formatting are preserved
3. **Modify Template Content**:
- **Frontmatter Section**:
- Set name: [skill-name] (kebab-case matching directory)
- Write description with clear "Use when..." clause for invocation triggers
- Add allowed-tools if tool restrictions specified in inputs
- Replace [Skill Name] placeholder with the actual skill name
- Customize the introduction section with specific purpose and context
- Define skill input/output specifications based on requirements
- Create ASCII skill diagram following the template pattern
- Implement skill steps using the template's phase structure
- Format subagent instructions using template's >>> <<< delimiters
- Apply skill path guidance from Phase 1 planning
4. **Clean & Finalize**:
- Remove all HTML comments containing "INSTRUCTION" markers
- Remove template placeholder instructions and guidance text
- Keep all skill content and user-facing documentation intact
- Ensure final document is clean and professional
- Verify SKILL.md is uppercase (not skill.md)
**Report**
**[IMPORTANT]** You MUST return the following execution report (<1000 tokens):
```yaml
status: success|failure|partial
summary: 'Brief description of skill creation completion'
modifications: ['[skill-name]/SKILL.md', ...]
outputs:
implementation_summary: 'Brief description of skill implementation approach'
frontmatter_valid: true|false
template_compliance: true|false
directory_created: true|false
issues: ['issue1', 'issue2', ...] # only if problems encountered
```
<<<
Step Configuration:
What You Send to Subagent:
In a single message, You assign the skill validation task to a different specialist subagent.
Request the subagent to perform the following validation review:
>>>
**ultrathink: adopt the Quality Assurance Specialist mindset**
- You're a **Quality Assurance Specialist** with expertise in skill documentation who follows these principles:
- **Template Compliance**: Verify exact adherence to template structure
- **Frontmatter Validation**: Ensure proper YAML structure and required fields
- **Content Quality**: Assess clarity, completeness, and professionalism
- **Logical Flow**: Ensure skill steps are sound and achievable
- **Documentation Standards**: Check formatting and consistency
<IMPORTANT>
You've to perform the task yourself. You CANNOT further delegate the work to another subagent
</IMPORTANT>
**Review Assignment**
You're assigned to validate the skill file that was created:
- **Skill File**: [skill file path from Step 2]
- **Implementation Summary**: [summary from Step 2]
- **Template Reference**: template:skill
**Review Steps**
1. **Read Created Skill**:
- Read the created skill file completely
- Compare against the template structure section by section
- Identify any missing or malformed sections
2. **Validate Frontmatter**:
- Check YAML frontmatter is present and properly formatted
- Verify name field matches directory name (kebab-case)
- Verify description includes "Use when..." clause for invocation
- Check allowed-tools if specified
- Ensure frontmatter delimiter (---) is correct
3. **Validate Template Compliance**:
- Check all required sections are present and properly structured
- Verify ASCII diagrams are properly formatted
- Confirm subagent instruction blocks follow template formatting
- Ensure all placeholder content has been replaced appropriately
4. **Assess Content Quality**:
- Review skill logic for soundness and clarity
- Check that inputs/outputs are well-defined
- Verify skill steps achieve the stated purpose
- Ensure professional documentation standards
5. **Validate Directory Structure**:
- Confirm skill lives in [plugin]/skills/[skill-name]/ directory
- Verify file is named SKILL.md (uppercase)
- Check directory structure is correct
**Report**
**[IMPORTANT]** You MUST return the following review report (<500 tokens):
```yaml
status: pass|fail
summary: 'Brief validation summary'
checks:
frontmatter_structure: pass|fail
frontmatter_use_when_clause: pass|fail
template_compliance: pass|fail
content_quality: pass|fail
skill_logic: pass|fail
directory_structure: pass|fail
file_naming: pass|fail
fatals: ['issue1', 'issue2', ...] # Only critical blockers
warnings: ['warning1', 'warning2', ...] # Non-blocking issues
recommendation: proceed|retry|rollback
```
<<<
What You Do:
Report the skill output as specified:
skill: create-skill
status: completed
outputs:
skill_file: '[plugin]/skills/[skill-name]/SKILL.md'
implementation_summary: 'Brief description of skill implementation approach'
creation_report:
frontmatter_validation: passed
template_compliance: passed
content_customization: completed
instruction_cleanup: completed
directory_structure: created
validation_report:
frontmatter_structure: passed
use_when_clause: present
structure_review: passed
logic_validation: passed
documentation_standards: passed
file_naming: correct
summary: |
Successfully created skill '[skill-name]' with complete template
customization and validation. Skill is ready for autonomous invocation
and properly structured for Claude Code auto-discovery.
Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.
Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.
Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.