Skip to content

platxa/platxa-skill-generator

BranchesTags

Repository files navigation

Platxa Skill Generator

Autonomous skill creator for Claude Code CLI using multi-phase orchestrated workflows with Task tool subagents.

Created by: DJ Patel — Founder & CEO, Platxa | https://platxa.com

Based on: Anthropic's Agent Skills specification

Table of Contents

Skills Catalog

Ready-to-use skills for the community! Browse and install production-ready skills from our curated catalog.

Skill Description Install
code-documenter Generate docs for Python, JS/TS, Java, Go, Rust ./scripts/install-from-catalog.sh code-documenter
commit-message Generate conventional commit messages ./scripts/install-from-catalog.sh commit-message
test-generator Generate comprehensive unit tests ./scripts/install-from-catalog.sh test-generator
pr-description Generate PR descriptions from git history ./scripts/install-from-catalog.sh pr-description 
# Quick install
./scripts/install-from-catalog.sh --list        # See all skills
./scripts/install-from-catalog.sh code-documenter  # Install one
./scripts/install-from-catalog.sh --all         # Install all

See the full Skills Catalog for details and contribution guidelines.

Overview

Platxa Skill Generator creates production-ready Claude Code skills by orchestrating specialized subagents through a multi-phase workflow:

  1. Discovery - Research domain knowledge via web search and existing skill analysis
  2. Architecture - Design skill structure based on type (Builder/Guide/Automation/Analyzer/Validator)
  3. Generation - Create SKILL.md, scripts, and reference documentation
  4. Validation - Verify against Anthropic's spec with quality scoring

Who Is This For?

  • Developers who want to extend Claude Code with custom capabilities
  • Teams building specialized AI workflows for their domains
  • Organizations creating reusable skill libraries

Features

Feature Description
Autonomous Creation Minimal user input, maximum automation
Multi-Phase Workflow Specialized subagents for each phase
Five Skill Types Builder, Guide, Automation, Analyzer, Validator
Quality Gates Validation against Anthropic's Agent Skills specification
Token Budget Management SKILL.md optimized for context efficiency
Script Security Executable helpers with security checks
Web Research Automatic domain knowledge discovery
Template System Pre-built templates for each skill type

Prerequisites

Before installing, ensure you have:

Requirement Version Purpose
Claude Code CLI Latest Required runtime environment
Git 2.0+ Clone the repository
Bash 4.0+ Run validation scripts
Python 3.10+ Token counting utility (optional)

Verify Prerequisites

# Check Claude Code is installed
claude --version

# Check Git
git --version

# Check Bash
bash --version

# Check Python (optional)
python3 --version

Quick Start (5 Minutes)

Get your first skill generated in 5 minutes:

Step 1: Install the Skill (1 minute)

# Clone and install as user skill
git clone https://github.com/platxa/platxa-skill-generator.git
cp -r platxa-skill-generator ~/.claude/skills/

Step 2: Verify Installation (30 seconds)

# Check the skill is installed
ls ~/.claude/skills/platxa-skill-generator/SKILL.md

Expected output:

/home/user/.claude/skills/platxa-skill-generator/SKILL.md

Step 3: Generate Your First Skill (3 minutes)

In Claude Code, run:

User: /skill-generator
Assistant: What skill would you like to create?
User: A skill that formats JSON files

The generator will:

  1. Research JSON formatting best practices
  2. Design a "Builder" type skill
  3. Generate the skill files
  4. Validate quality (>= 7.0/10)
  5. Offer to install the new skill

Step 4: Use Your New Skill

User: /json-formatter

Congratulations! You've created your first Claude Code skill.

Installation

Option 1: User Skill (Recommended)

Available across all your projects:

# Clone the repository
git clone https://github.com/platxa/platxa-skill-generator.git

# Install to user skills directory
cp -r platxa-skill-generator ~/.claude/skills/

# Verify installation
ls ~/.claude/skills/platxa-skill-generator/

Option 2: Project Skill

Available only in a specific project:

# Navigate to your project
cd /path/to/your/project

# Create skills directory if needed
mkdir -p .claude/skills

# Clone and install
git clone https://github.com/platxa/platxa-skill-generator.git .claude/skills/platxa-skill-generator

Option 3: Direct Download

# Download latest release
curl -L https://github.com/platxa/platxa-skill-generator/archive/main.zip -o skill-generator.zip
unzip skill-generator.zip
mv platxa-skill-generator-main ~/.claude/skills/platxa-skill-generator

Verify Installation

Run the validation script:

cd ~/.claude/skills/platxa-skill-generator
./scripts/validate-all.sh

Expected output:

[OK] SKILL.md exists
[OK] Frontmatter valid
[OK] Structure valid
[OK] All validations passed

Usage

Basic Usage

User: /skill-generator
Assistant: What skill would you like to create?
User: [Describe the skill you want]

With Specific Requirements

User: /skill-generator
User: Create a skill that:
- Analyzes Python code for security issues
- Uses OWASP Top 10 as reference
- Outputs a detailed report

Workflow Phases

The generator progresses through these phases automatically:

Phase Duration What Happens
Discovery 30-60s Web research, existing skill analysis
Architecture 10-20s Skill type selection, structure design
Generation 20-40s SKILL.md, scripts, references creation
Validation 5-10s Quality scoring, spec compliance check
Installation 5s Copy to skills directory

Command Options

Command Description
/skill-generator Interactive skill creation
/skill-generator "description" Start with initial description

Skill Types

Choose the right type for your use case:

Type Purpose Best For Key Sections
Builder Create new things Code generators, scaffolders Workflow, Templates, Output Checklist
Guide Teach/explain Tutorials, how-tos Steps, Examples, Best Practices
Automation Automate tasks CI/CD, batch processing Triggers, Scripts, Verification
Analyzer Inspect/audit Code review, security Checklist, Metrics, Reports
Validator Verify quality Testing, compliance Rules, Thresholds, Pass/Fail Criteria

Choosing a Skill Type

Is the primary goal to CREATE something new?
  └─ Yes → Builder

Is the primary goal to TEACH or EXPLAIN?
  └─ Yes → Guide

Is the primary goal to AUTOMATE a process?
  └─ Yes → Automation

Is the primary goal to INSPECT or AUDIT?
  └─ Yes → Analyzer

Is the primary goal to VERIFY or VALIDATE?
  └─ Yes → Validator

Examples

Example 1: Code Documenter (Automation)

User: /skill-generator
Assistant: What skill would you like to create?
User: A skill that generates documentation for Python code

[Discovery] Researching Python docstring standards...
  - Found: NumPy style, Google style, Sphinx format
  - Analyzing existing documentation tools

[Architecture] Designing skill structure...
  - Type: Automation
  - Scripts: generate-docs.sh, validate-docs.sh
  - References: docstring-styles.md, examples.md

[Generation] Creating skill files...
  - py-documenter/SKILL.md (142 lines)
  - py-documenter/scripts/generate-docs.sh
  - py-documenter/references/docstring-styles.md

[Validation] Quality check...
  - Frontmatter: OK
  - Structure: OK
  - Token budget: OK (2,847 tokens)
  - Quality score: 8.2/10

Install to ~/.claude/skills/py-documenter? (y/n)

Example 2: Security Analyzer (Analyzer)

User: /skill-generator
Assistant: What skill would you like to create?
User: A skill that scans code for security vulnerabilities

[Discovery] Researching security analysis...
  - Found: OWASP Top 10, CWE database, SAST patterns
  - Analyzing: semgrep rules, bandit patterns

[Architecture] Designing skill structure...
  - Type: Analyzer
  - References: owasp-checklist.md, vulnerability-patterns.md
  - Scripts: scan.sh, report.sh

[Generation] Creating skill files...
  - security-analyzer/SKILL.md (198 lines)
  - security-analyzer/references/owasp-checklist.md
  - security-analyzer/scripts/scan.sh

[Validation] Quality score: 7.8/10

Install to ~/.claude/skills/security-analyzer? (y/n)

Example 3: API Generator (Builder)

User: /skill-generator
Assistant: What skill would you like to create?
User: A skill that generates REST API endpoints from OpenAPI specs

[Discovery] Researching OpenAPI and API generation...
  - Found: OpenAPI 3.0 spec, code generation patterns
  - Analyzing: FastAPI, Express.js patterns

[Architecture] Designing skill structure...
  - Type: Builder
  - Templates: endpoint-template.py, router-template.py
  - Scripts: generate.sh, validate-openapi.sh

[Validation] Quality score: 8.5/10

Install to ~/.claude/skills/api-generator? (y/n)

Project Structure

platxa-skill-generator/
├── SKILL.md                 # Main skill definition (entry point)
├── README.md                # This file
├── LICENSE                  # MIT License
├── CONTRIBUTING.md          # Contribution guidelines
│
├── scripts/                 # Executable helpers
│   ├── validate-all.sh      # Run all validations
│   ├── validate-structure.sh
│   ├── validate-frontmatter.sh
│   ├── install-skill.sh     # Installation helper
│   ├── security-check.sh    # Security validation
│   └── count-tokens.py      # Token counting utility
│
├── references/              # Domain knowledge and patterns
│   ├── agents/              # Subagent definitions (7 agents)
│   │   ├── discovery-agent.md
│   │   ├── architecture-agent.md
│   │   ├── generation-agent.md
│   │   ├── validation-agent.md
│   │   ├── script-agent.md
│   │   ├── reference-agent.md
│   │   └── quality-agent.md
│   │
│   ├── patterns/            # Implementation patterns (40+ files)
│   │   ├── skill-types.md
│   │   ├── domain-discovery.md
│   │   ├── quality-criteria.md
│   │   ├── state-machine.md
│   │   └── ...
│   │
│   ├── templates/           # Skill templates by type
│   │   ├── builder-template.md
│   │   ├── guide-template.md
│   │   ├── automation-template.md
│   │   ├── analyzer-template.md
│   │   └── validator-template.md
│   │
│   └── state-schema.md      # State management schema
│
└── assets/                  # Static resources
    └── skill-template/      # Empty skill template
        ├── SKILL.md
        ├── scripts/.gitkeep
        └── references/.gitkeep

Quality Standards

Generated skills must meet these criteria to pass validation:

Required Checks

Check Requirement Validation
SKILL.md Exists with valid YAML frontmatter validate-frontmatter.sh
Name Hyphen-case, <= 64 characters Automated
Description <= 1024 characters Automated
Token Budget < 500 lines in SKILL.md count-tokens.py
Tools Only required tools listed Manual review
Scripts Executable and secure security-check.sh
References Real domain expertise Quality scoring
Quality Score >= 7.0/10 Quality agent

Quality Score Breakdown

Component Weight Criteria
Structure 20% Valid frontmatter, correct directories
Content 30% Clear instructions, useful examples
Expertise 25% Real domain knowledge in references
Security 15% Safe scripts, no vulnerabilities
Efficiency 10% Token budget, concise content

Troubleshooting

Common Issues

Skill not recognized after installation

Symptom: /skill-generator doesn't work

Solution:

# Verify the skill is in the correct location
ls ~/.claude/skills/platxa-skill-generator/SKILL.md

# Check frontmatter is valid
head -20 ~/.claude/skills/platxa-skill-generator/SKILL.md

# Restart Claude Code to reload skills

Quality score too low

Symptom: Generated skill fails validation with score < 7.0

Solution:

  • Provide more specific description of what the skill should do
  • Include target users and use cases
  • Let the generator complete all phases before interrupting

Scripts not executable

Symptom: Permission denied when running scripts

Solution:

# Make all scripts executable
chmod +x ~/.claude/skills/platxa-skill-generator/scripts/*.sh

Web research fails

Symptom: Discovery phase times out or returns empty

Solution:

  • Check internet connectivity
  • Try with a more specific or different topic
  • The generator will proceed with reduced quality if research fails

Getting Help

  1. Check existing issues: GitHub Issues
  2. Read the patterns: references/patterns/ contains detailed documentation
  3. Run diagnostics: ./scripts/validate-all.sh

Contributing

We welcome contributions! See CONTRIBUTING.md for:

  • Reporting issues
  • Submitting pull requests
  • Adding new skill types
  • Improving patterns and templates
  • Writing documentation

Quick Contribution Guide

# Fork the repository
# Clone your fork
git clone https://github.com/YOUR-USERNAME/platxa-skill-generator.git

# Create a feature branch
git checkout -b feature/your-feature

# Make changes and test
./scripts/validate-all.sh

# Commit and push
git add .
git commit -m "feat: add your feature"
git push origin feature/your-feature

# Open a Pull Request

License

MIT License - See LICENSE for details.

Acknowledgments

  • Anthropic - For Claude Code and the Agent Skills specification

Additional Resources

Version: 1.0.0 Created by: DJ Patel — Founder & CEO, Platxa | https://platxa.com Based on: Anthropic's Agent Skills Specification

About

Platxa Skill Generator

Resources

Readme

License

MIT license

Contributing

Contributing
Activity
Custom properties

Stars

4 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages