Skip to content

Agentic-Assets/Agent-Skills

BranchesTags

Repository files navigation

Version License Claude Code

Typing SVG

🎯 18 Skills • 🚀 9 Workflows • 🧠 Context Engineering • 📖 Progressive Disclosure

Overview

This is a personal collection of specialized Claude Code skills organized around PhD research, finance, real estate, and AI/ML development:

  • 📚 PhD Academic Business Research: Econometric analysis (PyFixest, STATA), data manipulation, publication-ready output
  • 💰 Financial Analysis & Services: WRDS data extraction, financial data analysis, portfolio analytics
  • 🏢 Real Estate (Residential/Commercial): Investment analysis, DCF/IRR modeling, institutional underwriting
  • 🤖 AI/ML/AI Agents: ML pipelines, prompt engineering, AI tool integration via MCP
  • 💻 Development & Technical Writing: FastAPI APIs, code quality, documentation, automation

Quick Start

Installation

# Copy skills to Claude Code directory
cp -r ./skills/* ~/.claude/skills/

Usage

Skills activate automatically based on your requests:

# Python API Development
"Build a FastAPI endpoint with async SQLAlchemy"
→ Activates: fastapi-expert

# Data Analysis
"Clean this DataFrame and create summary statistics"
→ Activates: pandas-pro

# Research
"Generate LaTeX tables from PyFixest models"
→ Activates: pyfixest-latex

# Debugging
"Help me debug this stack trace"
→ Activates: debugging-wizard

# Context Engineering
"Audit my CLAUDE.md and agent context setup"
→ Activates: context-engineering

Skills Overview

18 specialized skills organized by domain:

📚 PhD Academic Business Research (4 skills)

  • academic-writing: LaTeX manuscript drafting and revision for finance, economics, and real estate research
  • pyfixest-latex: PyFixest econometric models to publication-quality LaTeX (DiD, event studies, panel regression)
  • stata-accounting-research: STATA code patterns from published accounting research (entropy balancing, PSM, DiD, RDD, IV)
  • pandas-pro: DataFrame manipulation, data cleaning, aggregation, time series analysis for research

💰 Financial Analysis & Services (3 skills)

  • wrds-data-pull: WRDS data extraction (Compustat, CRSP, IBES, Thomson Reuters, BoardEx, ISS, CoreLogic, ZTRAX, CoStar)
  • pandas-pro: Financial data analysis, portfolio analytics, return calculations, risk metrics
  • pyfixest-latex: Financial econometric analysis and publication-ready output

🏢 Real Estate (Residential/Commercial) (1 skill)

  • cre-investment-analysis: Commercial real estate investment analysis, DCF/IRR modeling, business plans, institutional underwriting

🤖 AI/ML/AI Agents (3 skills)

  • ml-pipeline: ML pipelines with MLflow/Kubeflow, experiment tracking, feature stores, model lifecycle
  • prompt-engineer: LLM prompt design, chain-of-thought, few-shot learning, evaluation frameworks
  • mcp-developer: Model Context Protocol servers/clients for AI tool integration

💻 Development & Technical Writing (5 skills)

  • fastapi-expert: Async Python APIs with FastAPI, Pydantic V2, async SQLAlchemy, JWT auth
  • code-documenter: Docstrings, API docs (OpenAPI/Swagger), documentation sites, user guides
  • code-reviewer: PR reviews, code quality audits, security checks, refactoring suggestions
  • debugging-wizard: Systematic debugging, error investigation, root cause analysis
  • n8n-skills: n8n workflow automation, node configuration, workflow patterns

📊 Data Visualization (2 skills)

  • matplotlib: Low-level plotting library for full customization, novel plot types, fine-grained control
  • scientific-visualization: Publication-ready multi-panel figures with journal-specific formatting (Nature, Science, Cell)

🧠 Context Engineering (1 skill)

  • context-engineering: Audit, optimize, and architect the AI agent context layer (CLAUDE.md, hooks, commands, skills)

See SKILLS_GUIDE.md for when to use each skill, workflows, and examples.

Recommended Complementary Skills

These skills work seamlessly with the collection above. Highly recommended for document processing in research and real estate workflows:

Official Anthropic Document Skills

  • pdf - Extract text/tables from PDFs, OCR scanned documents

    • Essential for: Offering memos, appraisals, research papers, reports

  • xlsx - Read/write Excel spreadsheets, create financial models

    • Essential for: Rent rolls, operating statements, financial models, data analysis
    • Critical for CRE: Creates professional models with proper formulas and formatting

  • docx - Process Word documents, create reports

    • Essential for: Investment memos, business plans, market reports

  • pptx - Create/analyze PowerPoint presentations

    • Essential for: Investment committee presentations, board decks

Installation: These skills are built-in to Claude.ai and available at /mnt/skills/public/ in Claude Code.

Integration Example:

pdf (extract offering memo) → pandas-pro (analyze data) →
cre-investment-analysis (perform analysis) → xlsx (create model) →
pptx (create presentation)

Discover More Skills

For a comprehensive, up-to-date catalog of available skills:

🔗 skills.sh - Browse hundreds of community and official skills

Architecture

Progressive Disclosure Pattern

Each skill follows this structure:

skills/fastapi-expert/
├── SKILL.md                    # Lean core (~80 lines)
│   ├── Role definition
│   ├── When to use
│   ├── Core workflow
│   └── Routing table          # Points to references
└── references/                 # Loaded on-demand
    ├── async-patterns.md
    ├── pydantic-v2.md
    ├── authentication.md
    └── ...

How It Works:

  1. Skill loads with minimal context (~80 lines)
  2. Claude reads the routing table
  3. Loads specific references only when context requires
  4. 50% faster initial responses, surgical precision when needed

Stats:

  • 18 skills
  • 91 reference files
  • ~50% token reduction through progressive disclosure

Usage Patterns

Multi-Skill Workflows

Complex tasks combine multiple skills:

Academic Research Paper:

wrds-data-pull → pandas-pro → pyfixest-latex → code-documenter → code-reviewer

Financial Analysis Project:

wrds-data-pull → pandas-pro → code-documenter → code-reviewer

Real Estate Investment Analysis:

pandas-pro → cre-investment-analysis → code-documenter

ML/AI Pipeline:

pandas-pro → ml-pipeline → prompt-engineer → mcp-developer → code-documenter

API Development:

fastapi-expert → debugging-wizard → code-documenter → code-reviewer

Accounting Research (STATA):

wrds-data-pull → stata-accounting-research → code-documenter

Tech Stack Coverage

Research & Academia

  • Econometrics: PyFixest (DiD, event studies, panel regression), STATA (PSM, entropy balancing, RDD, IV)
  • Data Sources: WRDS (Compustat, CRSP, IBES, Thomson Reuters, BoardEx, ISS, CoreLogic, ZTRAX)
  • Output: LaTeX tables/figures, publication-quality plots
  • Languages: Python 3.11+, STATA 18

Financial Analysis

  • Data: pandas, NumPy, SciPy
  • Sources: WRDS financial databases
  • Methods: Portfolio analytics, return calculations, risk metrics, event studies
  • Integration: SQL queries, CUSIP/GVKEY/PERMNO linking

Real Estate

  • Analysis: DCF modeling, IRR analysis, NOI calculations, cap rate analysis
  • Property Types: Multifamily, office, retail, industrial, mixed-use
  • Output: Investment memos, underwriting models, feasibility studies

AI/ML

  • Pipelines: MLflow, Kubeflow, experiment tracking, feature stores
  • LLMs: Prompt engineering, chain-of-thought, few-shot learning
  • Integration: Model Context Protocol (MCP), AI tool development
  • ML Stack: scikit-learn, PyTorch, TensorFlow

Development

  • Backend: FastAPI, Pydantic V2, async SQLAlchemy, OpenAPI/Swagger
  • Automation: n8n workflow automation, webhook processing
  • Documentation: Docstrings, API docs, technical writing
  • Quality: Code review, debugging, security audits

Project Structure

Agent-Skills/
├── .claude-plugin/
│   ├── plugin.json           # Plugin metadata
│   └── marketplace.json      # Marketplace configuration
├── skills/                   # 17 specialized skills
│   ├── academic-writing/
│   ├── code-documenter/
│   ├── code-reviewer/
│   ├── context-engineering/
│   ├── cre-investment-analysis/
│   ├── debugging-wizard/
│   ├── fastapi-expert/
│   ├── matplotlib/
│   ├── mcp-developer/
│   ├── ml-pipeline/
│   ├── n8n-skills/
│   ├── pandas-pro/
│   ├── prompt-engineer/
│   ├── pyfixest-latex/
│   ├── scientific-visualization/
│   ├── stata-accounting-research/
│   └── wrds-data-pull/
├── commands/                 # 9 workflow commands
│   ├── common-ground/
│   └── project/
├── scripts/
│   ├── update-docs.py        # Update version and counts
│   └── validate-skills.py    # Validate skill integrity
├── docs/
│   ├── ATLASSIAN_MCP_SETUP.md
│   ├── COMMON_GROUND.md
│   └── WORKFLOW_COMMANDS.md
├── README.md
├── SKILLS_GUIDE.md          # Quick reference guide
├── CLAUDE.md                # Project configuration
└── CONTRIBUTING.md          # Contribution guidelines

Documentation

  • SKILLS_GUIDE.md - Quick reference for when to use each skill
  • CLAUDE.md - Project configuration and skill authorship standards
  • CONTRIBUTING.md - Guidelines for contributing
  • skills/*/SKILL.md - Individual skill documentation
  • skills/*/references/ - Deep-dive reference materials

Maintenance

Update Documentation

When adding/removing skills or changing versions:

# Update version in version.json, then:
python scripts/update-docs.py

# Validate all skills
python scripts/validate-skills.py

Validate Skills

Before committing changes:

# Run full validation
python scripts/validate-skills.py

# Validate specific skill
python scripts/validate-skills.py --skill fastapi-expert

# Check YAML only
python scripts/validate-skills.py --check yaml

Contributing

This is a personal skill collection, but contributions are welcome!

Adding a New Skill

  1. Create skill directory:

    mkdir -p skills/my-skill/references

  2. Create lean SKILL.md with YAML frontmatter:

    ---
name: my-skill
description: Use when [triggering conditions]
triggers:
  - keyword1
  - keyword2
role: specialist|expert|architect
scope: implementation|review|design
output-format: code|document|report
---

  3. Create reference files (100-600 lines each)

  4. Update version.json and run:

    python scripts/update-docs.py
python scripts/validate-skills.py

  5. Test locally and commit

See CLAUDE.md for detailed authorship standards.

Changelog

See CHANGELOG.md for version history and release notes.

License

MIT License - See LICENSE file for details.

Original template: jeffallan/claude-skills

Skill Sources & Attribution

Some skills in this collection were created by other contributors in the community:

Community-Contributed Skills

Related Resources

Check for Updates: Visit the original repositories above to get the latest versions and contribute back improvements.

About

Personal skill collection for Python development, data analysis, econometric research, and AI integration.

Built on the Agent Skills specification.

Originally forked from jeffallan/claude-skills.

Built for Claude Code | 9 Workflows | 91 Reference Files | 18 Skills

About

No description, website, or topics provided.

Resources

Readme

License

MIT license

Contributing

Contributing
Activity
Custom properties

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages