Chapter 1: Prompt Template System

Template Architecture

The Prompt Engineering Kit provides eight structured prompt templates

organized by reasoning strategy. Each template is a Jinja-like text file

with {placeholder} variables, chain-of-thought scaffolding, and optional

few-shot examples.

Template Structure

templates/
├── zero-shot/
│   ├── summarize.txt
│   └── classify.txt
├── few-shot/
│   ├── code-review.txt
│   └── extract-entities.txt
├── chain-of-thought/
│   ├── solve-math.txt
│   └── debug-reasoning.txt
├── role-play/
│   └── interview-trainer.txt
└── structured-output/
    └── json-extract.txt

Building a Prompt

python src/prompt_engineering_kit.py --build code-review --vars \
  '{"language":"python","context":"auth microservice","focus":"security"}'

This produces a formatted prompt ready to paste into any LLM chat interface:

You are a senior engineer reviewing Python code for an auth microservice.
Focus specifically on security vulnerabilities.

Code:
[code to review]

Checklist:
- Injection vulnerabilities
- Authentication bypasses
- Hardcoded credentials
- Input validation gaps
- Proper error handling

Variable Injection

The kit validates all placeholders at build time. Missing variables produce

a clear error listing exactly which template fields need values:

python src/prompt_engineering_kit.py --build summarize --vars '{}'
# Error: Template 'summarize' requires: text, length, style

Prompt Scoring

The quality analyzer rates prompts on a 0-100 scale across four dimensions:

python src/prompt_engineering_kit.py --score "Explain quantum computing simply"
# Length:  42/100  (too short — expand the prompt)
# Clarity: 78/100  (decent, but define "quantum computing")
# Structure: 35/100 (add sections for depth, audience, format)
# Specificity: 12/100 (no constraints on format or depth)

For multi-step reasoning chains, see 02_chain-of-thought-templates.md.

Project Structure

Follow this guide to get Prompt Engineering Kit up and running in your environment.

Project Structure

prompt-engineering-kit/
├── README.md
├── LICENSE
├── src/
│   └── prompt_engineering_kit.py    # Core engine (~400 lines)
└── examples/
    ├── basic_usage.py               # Programmatic usage example
    └── custom_templates.json        # Sample custom template library

CLI Reference