@intentsolutionsio/ai-ml-engineering-pack 1.0.0 → 1.0.4

package/README.md

@@ -8,31 +8,35 @@ Master prompt engineering, LLM integration, RAG systems, and AI safety with 12 s
 [![Version](https://img.shields.io/badge/version-1.0.0-blue.svg)](https://github.com/jeremylongshore/claude-code-plugins)
 [![Claude Code](https://img.shields.io/badge/Claude_Code-0.1.0+-purple.svg)](https://claude.ai/code)
 
-##  What's Included
+## What's Included
 
 **12 specialized plugins across 4 AI/ML categories:**
 
 ### 1. Prompt Engineering (3 plugins)
+
 - **prompt-architect** (agent) - Expert in CoT reasoning, few-shot learning, and advanced prompt patterns
 - **prompt-optimizer** (agent) - Reduce LLM costs by 60-90% while maintaining quality
 - **prompt-template-gen** (command: `/ptg`) - Generate production-ready prompt templates with type safety
 
 ### 2. LLM Integration (3 plugins)
+
 - **llm-integration-expert** (agent) - Production API patterns, error handling, streaming, rate limiting
 - **model-selector** (agent) - Choose optimal models based on cost, quality, latency requirements
 - **llm-api-scaffold** (command: `/las`) - Generate complete LLM API with FastAPI, Docker, monitoring
 
 ### 3. RAG Systems (3 plugins)
+
 - **rag-architect** (agent) - Design RAG systems, chunking strategies, retrieval optimization
 - **vector-db-expert** (agent) - Select and configure vector databases (Pinecone, Qdrant, Weaviate, etc.)
 - **rag-pipeline-gen** (command: `/rpg`) - Generate complete RAG pipeline with embeddings and retrieval
 
 ### 4. AI Safety (3 plugins)
+
 - **ai-safety-expert** (agent) - Content filtering, PII detection, bias mitigation, compliance
 - **prompt-injection-defender** (agent) - Defend against prompt injection and jailbreak attacks
 - **ai-monitoring-setup** (command: `/ams`) - Set up LLM monitoring, cost tracking, and alerts
 
-##  Quick Start
+## Quick Start
 
 ### Installation
 
@@ -47,7 +51,7 @@ claude plugin install ai-ml-engineering-pack@claude-code-plugins-plus
 claude plugin list
 ```
 
-**Full installation guide:** [INSTALLATION.md](./INSTALLATION.md)
+**Full installation guide:** INSTALLATION.md
 
 ### 10-Minute Tutorial
 
@@ -75,9 +79,9 @@ claude
 "Implement PII detection and toxicity filtering for my chatbot"
 ```
 
-**Complete tutorial:** [QUICK_START.md](./QUICK_START.md)
+**Complete tutorial:** QUICK_START.md
 
-##  ROI & Value Proposition
+## ROI & Value Proposition
 
 Real-world results from production deployments:
 
@@ -92,16 +96,18 @@ Real-world results from production deployments:
 
 **Average ROI: 29,351%** | **Average payback period: 3 days**
 
-**Detailed case studies:** [USE_CASES.md](./USE_CASES.md)
+**Detailed case studies:** USE_CASES.md
 
-##  Plugin Reference
+## Plugin Reference
 
 ### Prompt Engineering
 
 #### `prompt-architect` (Agent)
+
 Expert in advanced prompt engineering techniques and patterns.
 
 **Capabilities:**
+
 - Chain-of-Thought (CoT) reasoning
 - Few-shot and zero-shot learning
 - Prompt composition patterns
@@ -109,6 +115,7 @@ Expert in advanced prompt engineering techniques and patterns.
 - Multi-modal prompts (text + images)
 
 **When to use:**
+
 - "Design a prompt for [complex task]"
 - "Improve this prompt: [existing prompt]"
 - "What's the best prompting technique for [use case]?"
@@ -118,9 +125,11 @@ Expert in advanced prompt engineering techniques and patterns.
 ---
 
 #### `prompt-optimizer` (Agent)
+
 Optimize prompts for cost reduction (60-90% savings) while maintaining quality.
 
 **Capabilities:**
+
 - Token reduction techniques (remove verbosity, use abbreviations)
 - Prompt caching strategies
 - Model selection guidance (cheap vs expensive)
@@ -128,11 +137,13 @@ Optimize prompts for cost reduction (60-90% savings) while maintaining quality.
 - ROI calculation
 
 **When to use:**
+
 - "Reduce the cost of this prompt: [prompt]"
 - "Optimize my prompts for $1000/month budget"
 - "How can I reduce token usage by 70%?"
 
 **Example:**
+
 ```
 Before (52 tokens): "I would like you to please analyze..."
 After (15 tokens): "Analyze and summarize main points."
@@ -144,9 +155,11 @@ Savings: 71% token reduction = $0.15/1000 calls (GPT-4)
 ---
 
 #### `/ptg` - Prompt Template Generator (Command)
+
 Generate production-ready prompt templates with type safety and validation.
 
 **Usage:**
+
 ```bash
 /ptg
 
@@ -158,6 +171,7 @@ Generate production-ready prompt templates with type safety and validation.
 ```
 
 **Generated output:**
+
 - Python: Pydantic models with type safety
 - TypeScript: Zod schemas with validation
 - Usage examples
@@ -165,6 +179,7 @@ Generate production-ready prompt templates with type safety and validation.
 - Unit tests
 
 **Example output:**
+
 ```python
 @dataclass
 class ProductDescriptionInput:
@@ -186,9 +201,11 @@ class ProductDescriptionGenerator:
 ### LLM Integration
 
 #### `llm-integration-expert` (Agent)
+
 Production patterns for LLM API integration with error handling and reliability.
 
 **Capabilities:**
+
 - Multi-provider integration (OpenAI, Anthropic, Google, Cohere)
 - Exponential backoff retry logic
 - Rate limiting (token bucket, sliding window)
@@ -198,11 +215,13 @@ Production patterns for LLM API integration with error handling and reliability.
 - Token counting and cost tracking
 
 **When to use:**
+
 - "Implement LLM API integration with retry logic"
 - "Add streaming support to my chatbot"
 - "Build multi-provider fallback system"
 
 **Code examples:**
+
 ```python
 # Retry with exponential backoff
 @retry_with_backoff(max_retries=3, base_delay=1.0)
@@ -219,9 +238,11 @@ await rate_limiter.wait_for_token()
 ---
 
 #### `model-selector` (Agent)
+
 Guide model selection based on cost, quality, latency, and use case requirements.
 
 **Capabilities:**
+
 - Model comparison matrix (GPT-4, Claude 3, Gemini)
 - Pricing analysis (per 1M tokens)
 - Latency benchmarks
@@ -230,11 +251,13 @@ Guide model selection based on cost, quality, latency, and use case requirements
 - A/B testing frameworks
 
 **When to use:**
+
 - "Which model should I use for customer support?"
 - "Compare GPT-4 vs Claude 3 Opus for code generation"
 - "How can I reduce costs with model cascading?"
 
 **Model comparison:**
+
 | Model | Input ($/1M) | Output ($/1M) | Latency | Best For |
 |-------|-------------|---------------|---------|----------|
 | GPT-4 Turbo | $10 | $30 | 3-5s | Complex reasoning |
@@ -247,9 +270,11 @@ Guide model selection based on cost, quality, latency, and use case requirements
 ---
 
 #### `/las` - LLM API Scaffold (Command)
+
 Generate complete production-ready LLM API integration code.
 
 **Usage:**
+
 ```bash
 /las
 
@@ -261,6 +286,7 @@ Generate complete production-ready LLM API integration code.
 ```
 
 **Generated files:**
+
 ```
 llm-api/
 ├── main.py                 # FastAPI application
@@ -275,22 +301,25 @@ llm-api/
 ```
 
 **Features included:**
--  Exponential backoff retry (3 attempts)
--  Rate limiting (token bucket algorithm)
--  Response caching (Redis, 5 min TTL)
--  Streaming support (SSE)
--  Cost tracking
--  Prometheus metrics
--  Docker deployment
+
+- Exponential backoff retry (3 attempts)
+- Rate limiting (token bucket algorithm)
+- Response caching (Redis, 5 min TTL)
+- Streaming support (SSE)
+- Cost tracking
+- Prometheus metrics
+- Docker deployment
 
 ---
 
 ### RAG Systems
 
 #### `rag-architect` (Agent)
+
 Expert in designing and optimizing Retrieval-Augmented Generation systems.
 
 **Capabilities:**
+
 - RAG architecture patterns
 - Chunking strategies (fixed, recursive, semantic)
 - Embedding model selection
@@ -299,11 +328,13 @@ Expert in designing and optimizing Retrieval-Augmented Generation systems.
 - Evaluation metrics (MRR, NDCG)
 
 **When to use:**
+
 - "Design a RAG system for customer support knowledge base"
 - "What chunking strategy should I use for legal documents?"
 - "How can I improve retrieval accuracy?"
 
 **Chunking strategies:**
+
 ```python
 # Fixed-size (simple, fast)
 chunks = [text[i:i+512] for i in range(0, len(text), 512)]
@@ -324,9 +355,11 @@ chunks = semantic_splitter.split_by_meaning(text)
 ---
 
 #### `vector-db-expert` (Agent)
+
 Select and optimize vector databases for RAG systems.
 
 **Capabilities:**
+
 - Database comparison (Pinecone, Qdrant, Weaviate, ChromaDB, pgvector, Milvus)
 - HNSW index tuning
 - Scaling strategies (sharding, replication)
@@ -334,11 +367,13 @@ Select and optimize vector databases for RAG systems.
 - Migration planning
 
 **When to use:**
+
 - "Which vector database should I use for 10M documents?"
 - "How do I tune HNSW parameters for better performance?"
 - "Compare Pinecone vs Qdrant for my use case"
 
 **Database comparison:**
+
 | Database | Best For | Pricing | Hosting |
 |----------|---------|---------|---------|
 | Pinecone | Managed, auto-scaling | $0.096/GB/month | Cloud only |
@@ -352,9 +387,11 @@ Select and optimize vector databases for RAG systems.
 ---
 
 #### `/rpg` - RAG Pipeline Generator (Command)
+
 Generate complete RAG pipeline with all components.
 
 **Usage:**
+
 ```bash
 /rpg
 
@@ -367,6 +404,7 @@ Generate complete RAG pipeline with all components.
 ```
 
 **Generated files:**
+
 ```
 rag-system/
 ├── document_loader.py      # PDF/DOCX/TXT loaders
@@ -382,24 +420,27 @@ rag-system/
 ```
 
 **Features included:**
--  Multi-format document loading (PDF, DOCX, TXT, MD)
--  Recursive chunking (512 tokens, 50 overlap)
--  Vector similarity search
--  Cohere reranking (optional)
--  Source attribution with page numbers
--  Query expansion
--  Caching
--  FastAPI REST endpoints
--  Docker deployment
+
+- Multi-format document loading (PDF, DOCX, TXT, MD)
+- Recursive chunking (512 tokens, 50 overlap)
+- Vector similarity search
+- Cohere reranking (optional)
+- Source attribution with page numbers
+- Query expansion
+- Caching
+- FastAPI REST endpoints
+- Docker deployment
 
 ---
 
 ### AI Safety
 
 #### `ai-safety-expert` (Agent)
+
 Comprehensive AI safety with content filtering, PII protection, and bias mitigation.
 
 **Capabilities:**
+
 - Toxicity detection (BERT-based classification)
 - PII detection and redaction (Presidio)
 - Bias detection (gender, racial, age)
@@ -408,12 +449,14 @@ Comprehensive AI safety with content filtering, PII protection, and bias mitigat
 - GDPR/CCPA/HIPAA compliance
 
 **When to use:**
+
 - "Implement PII detection for user inputs"
 - "Add toxicity filtering to my chatbot"
 - "Detect and mitigate bias in LLM outputs"
 - "Ensure HIPAA compliance for medical data"
 
 **Safety pipeline:**
+
 ```python
 class SafetyGuardrails:
     async def safe_completion(self, user_input: str, llm):
@@ -434,6 +477,7 @@ class SafetyGuardrails:
 ```
 
 **PII detection:**
+
 - Email addresses, phone numbers, SSN
 - Credit card numbers
 - IP addresses
@@ -445,9 +489,11 @@ class SafetyGuardrails:
 ---
 
 #### `prompt-injection-defender` (Agent)
+
 Defend against prompt injection attacks and jailbreaks.
 
 **Capabilities:**
+
 - Pattern-based detection (regex for common attacks)
 - ML classification (fine-tuned BERT model)
 - Input sanitization
@@ -456,11 +502,13 @@ Defend against prompt injection attacks and jailbreaks.
 - Jailbreak detection (DAN, Developer Mode, etc.)
 
 **When to use:**
+
 - "Protect my chatbot from prompt injection"
 - "Detect jailbreak attempts"
 - "Validate user inputs for manipulation"
 
 **Attack patterns detected:**
+
 ```python
 ATTACK_PATTERNS = [
     r'ignore\s+(all\s+)?(previous|prior|above)\s+instructions',
@@ -472,6 +520,7 @@ ATTACK_PATTERNS = [
 ```
 
 **Defense strategies:**
+
 1. **Detection:** Identify attack patterns
 2. **Sanitization:** Remove/escape dangerous inputs
 3. **Validation:** Verify outputs don't leak system prompts
@@ -482,9 +531,11 @@ ATTACK_PATTERNS = [
 ---
 
 #### `/ams` - AI Monitoring Setup (Command)
+
 Set up comprehensive LLM monitoring with cost tracking and alerting.
 
 **Usage:**
+
 ```bash
 /ams
 
@@ -496,6 +547,7 @@ Set up comprehensive LLM monitoring with cost tracking and alerting.
 ```
 
 **Generated files:**
+
 ```
 monitoring/
 ├── metrics.py              # Prometheus metrics
@@ -508,6 +560,7 @@ monitoring/
 ```
 
 **Metrics collected:**
+
 - Request count (by model, status)
 - Latency (p50, p95, p99)
 - Token usage (input, output)
@@ -516,12 +569,14 @@ monitoring/
 - Cache hit rate
 
 **Alerts configured:**
+
 - Budget threshold (80%, 90%, 100%)
 - High error rate (>5%)
 - Slow responses (>10s)
 - Token limit approaching
 
 **Dashboards:**
+
 - Real-time request monitoring
 - Cost tracking (daily, weekly, monthly)
 - Model performance comparison
@@ -529,14 +584,14 @@ monitoring/
 
 ---
 
-##  Documentation
+## Documentation
 
-- **[Installation Guide](./INSTALLATION.md)** - Prerequisites, setup, verification
-- **[Quick Start](./QUICK_START.md)** - 10-minute tutorial with examples
-- **[Use Cases](./USE_CASES.md)** - Real-world applications with ROI
-- **[Troubleshooting](./000-docs/157-DR-FAQS-troubleshooting.md)** - Common issues and solutions
+- **Installation Guide** - Prerequisites, setup, verification
+- **Quick Start** - 10-minute tutorial with examples
+- **Use Cases** - Real-world applications with ROI
+- **Troubleshooting** - Common issues and solutions
 
-##  Example Workflows
+## Example Workflows
 
 ### Build a Customer Support Bot (10 minutes)
 
@@ -584,7 +639,7 @@ Use case: Customer feedback analysis
 
 ### Build RAG System for Legal Documents (15 minutes)
 
-```bash
+```text
 claude
 
 # 1. Design RAG architecture
@@ -608,42 +663,47 @@ Track: accuracy, retrieval time, cost per query
 
 **Result:** Legal document analysis system with 94% accuracy, 82ms latency, PII protection.
 
-##  Learning Resources
+## Learning Resources
 
 ### Video Tutorials (Coming Soon)
+
 - Prompt Engineering Masterclass (30 min)
 - Building Production RAG Systems (45 min)
 - AI Safety Best Practices (20 min)
 
 ### Blog Posts
+
 - [Reduce LLM Costs by 90%](https://example.com/reduce-llm-costs)
 - [Building RAG Systems That Actually Work](https://example.com/rag-systems)
 - [Comprehensive Guide to AI Safety](https://example.com/ai-safety)
 
 ### Community
+
 - [Discord](https://discord.com/invite/6PPFFzqPDZ) - #claude-code channel
 - [GitHub Discussions](https://github.com/jeremylongshore/claude-code-plugins/discussions)
 - [Stack Overflow](https://stackoverflow.com/questions/tagged/claude-code) - `claude-code` tag
 
-##  Pricing
+## Pricing
 
 **One-time purchase: $79**
 
 What's included:
--  All 12 plugins (lifetime access)
--  Free updates and new plugins
--  Email support
--  Community Discord access
--  Documentation and examples
+
+- All 12 plugins (lifetime access)
+- Free updates and new plugins
+- Email support
+- Community Discord access
+- Documentation and examples
 
 **Compare to alternatives:**
+
 - Manual implementation: 40+ hours ($4,000 at $100/hour)
 - Consultants: $150-300/hour × 40 hours = $6,000-12,000
 - AI/ML Engineering Pack: **$79** (99% cost savings)
 
 **Average payback period: 3 days**
 
-[Buy Now on Gumroad](https://gumroad.com/l/ai-ml-engineering-pack) | [Volume Licensing](mailto:[email protected])
+Buy Now on Gumroad | [Volume Licensing](mailto:[email protected])
 
 ## 🆘 Support
 
@@ -655,44 +715,47 @@ What's included:
 
 **Community:** Join Discord for community support
 
-##  Updates
+## Updates
 
 **Current version:** 1.0.0
 
 **Update policy:** Free updates for life, including new plugins and features
 
 **Changelog:**
+
 - **v1.0.0** (2025-10-10) - Initial release with 12 plugins
 
 To update:
+
 ```bash
 claude plugin update ai-ml-engineering-pack
 ```
 
-##  License
+## License
 
-MIT License - See [LICENSE](./000-docs/001-BL-LICN-license.txt) for details
+MIT License - See LICENSE for details
 
 **Commercial use permitted** - Use in commercial projects, redistribute, modify
 
-##  Acknowledgments
+## Acknowledgments
 
 Built with:
+
 - [Claude Code](https://claude.ai/code) - AI-powered development CLI
 - [LangChain](https://langchain.com) - LLM framework
 - [Presidio](https://microsoft.github.io/presidio/) - PII detection
 - [Qdrant](https://qdrant.tech) - Vector database
 - [FastAPI](https://fastapi.tiangolo.com) - Modern Python framework
 
-##  Ready to Get Started?
+## Ready to Get Started?
 
-1. **[Install the pack](./INSTALLATION.md)** - 5-minute setup
-2. **[Complete Quick Start](./QUICK_START.md)** - Build your first AI feature in 10 minutes
-3. **[Explore use cases](./USE_CASES.md)** - See real-world ROI examples
+1. **Install the pack** - 5-minute setup
+2. **Complete Quick Start** - Build your first AI feature in 10 minutes
+3. **Explore use cases** - See real-world ROI examples
 4. **[Join the community](https://discord.com/invite/6PPFFzqPDZ)** - Connect with other AI/ML engineers
 
 ---
 
 **Questions?** Email [email protected] or open a [GitHub issue](https://github.com/jeremylongshore/claude-code-plugins/issues).
 
-**Built by AI engineers, for AI engineers.** 
+**Built by AI engineers, for AI engineers.**

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@intentsolutionsio/ai-ml-engineering-pack",
-  "version": "1.0.0",
+  "version": "1.0.4",
   "description": "Professional AI/ML Engineering toolkit: Prompt engineering, LLM integration, RAG systems, AI safety with 12 expert plugins",
   "keywords": [
     "ai",

package/skills/optimizing-prompts/SKILL.md

@@ -1,13 +1,22 @@
 ---
 name: optimizing-prompts
-description: |
-  Execute this skill optimizes prompts for large language models (llms) to reduce token usage, lower costs, and improve performance. it analyzes the prompt, identifies areas for simplification and redundancy removal, and rewrites the prompt to be more conci... Use when optimizing performance. Trigger with phrases like 'optimize', 'performance', or 'speed up'.
+description: 'Execute this skill optimizes prompts for large language models (llms)
+  to reduce token usage, lower costs, and improve performance. it analyzes the prompt,
+  identifies areas for simplification and redundancy removal, and rewrites the prompt
+  to be more conci... Use when optimizing performance. Trigger with phrases like ''optimize'',
+  ''performance'', or ''speed up''.
+
+  '
 allowed-tools: Read, Write, Edit, Grep, Glob, Bash(cmd:*)
 version: 1.0.0
 author: Jeremy Longshore <jeremy@intentsolutions.io>
 license: MIT
-compatible-with: claude-code, codex, openclaw
-tags: [packages, llm, performance, cost-optimization]
+tags:
+- packages
+- llm
+- performance
+- cost-optimization
+compatibility: Designed for Claude Code, also compatible with Codex and OpenClaw
 ---
 # Ai Ml Engineering Pack
 
@@ -26,6 +35,7 @@ Refine prompts for optimal LLM performance. It streamlines prompts to minimize t
 ## When to Use This Skill
 
 This skill activates when you need to:
+
 - Reduce the cost of using an LLM.
 - Improve the speed of LLM responses.
 - Enhance the quality or clarity of LLM outputs by refining the prompt.
@@ -37,6 +47,7 @@ This skill activates when you need to:
 User request: "Optimize this prompt for cost and quality: 'I would like you to create a detailed product description for a new ergonomic office chair, highlighting its features, benefits, and target audience, and also include information about its warranty and return policy.'"
 
 The skill will:
+
 1. Analyze the prompt for redundancies and areas for simplification.
 2. Rewrite the prompt to be more concise: "Create a product description for an ergonomic office chair. Include features, benefits, target audience, warranty, and return policy."
 3. Provide the optimized prompt and explain the token reduction achieved.
@@ -46,6 +57,7 @@ The skill will:
 User request: "Optimize this prompt for better summarization: 'Please read the following document and provide a comprehensive summary of all the key points, main arguments, supporting evidence, and overall conclusion, ensuring that the summary is accurate, concise, and easy to understand.'"
 
 The skill will:
+
 1. Identify areas for improvement in the prompt's clarity and focus.
 2. Rewrite the prompt to be more direct: "Summarize this document, including key points, arguments, evidence, and the conclusion."
 3. Present the optimized prompt and explain how it enhances summarization performance.
@@ -85,4 +97,4 @@ The skill produces structured output relevant to the task.
 ## Resources
 
 - Project documentation
-- Related skills and commands
+- Related skills and commands

package/skills/optimizing-prompts/assets/example_prompts.md

@@ -243,4 +243,4 @@ Output: A boolean value indicating whether a prompt injection attempt is detecte
 Implement heuristics and/or machine learning models to identify suspicious patterns.  Check for phrases like "ignore previous instructions" and "as an AI language model".
 ```
 
-Remember to adapt these examples to your specific needs and experiment with different prompts to achieve the best results. Good luck!
+Remember to adapt these examples to your specific needs and experiment with different prompts to achieve the best results. Good luck!

package/skills/optimizing-prompts/assets/optimization_report_template.md

@@ -101,4 +101,4 @@
 
 ### 8.3. Data Sources
 
-`[Describe the data sources used to evaluate the prompt's performance.  Specify the size and characteristics of the dataset.]`
+`[Describe the data sources used to evaluate the prompt's performance.  Specify the size and characteristics of the dataset.]`

package/skills/optimizing-prompts/references/README.md

@@ -1,4 +1,3 @@
 # References
 
 Bundled resources for ai-ml-engineering-pack skill
-

package/skills/optimizing-prompts/scripts/README.md

@@ -6,6 +6,6 @@ Bundled resources for ai-ml-engineering-pack skill
 - [x] prompt_validator.py: Script to validate prompt syntax and structure.
 - [x] cost_estimator.py: Script to estimate the cost of running a prompt on different LLMs.
 
-
 ## Auto-Generated
+
 Scripts generated on 2025-12-10 03:48:17