skills.rest
cjennings
cjenningsCOMMUNITY
Explore all skills by this author →
Website𝕏@craigjenningsGitHub

prompt-engineering

Classify prompt types and apply persuasion frameworks to reduce failure modes.

Updated Aug 27, 2026
Search Tags:#llm#frontmatter#prompt-engineering#system-prompts#critique-mode#design-mode#persuasion-framework
One-click install
npx skills add https://github.com/cjennings/rulesets --skill prompt-engineering-cjennings
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: prompt-engineering
Source: https://github.com/cjennings/rulesets/tree/main/prompt-engineering
Command: npx skills add https://github.com/cjennings/rulesets --skill prompt-engineering-cjennings

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Craft prompts that resist common failure modes and drive reliable, predictable behavior across design and critique workflows. This Skill helps you classify prompt types, apply persuasion frameworks appropriately, manage degrees of freedom, and keep context-window economics in check.

Core Features & Use Cases

  • Classifies prompts into Discipline-enforcing, Guidance, Collaborative, and Reference types to tailor behavior.
  • Applies a deliberate persuasion framework aligned with the prompt type to reduce undesired outputs.
  • Advises on choosing degrees of freedom (low/medium/high) based on task fragility and context window economics.
  • Provides dual workflow modes: design mode for creating new prompts and critique mode for rewriting drafts to resist common failure modes.
  • Includes an ethics test and anti-pattern guidance to ensure safe, trustworthy prompts.

Quick Start

Draft a new system prompt for a hypothetical data-privacy check by classifying its type, selecting persuasion principles, and producing a concise, deterministic prompt that requires no exceptions.

Frequently Asked Questions about prompt-engineering

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I design system prompts that resist common failure modes?▼

To design system prompts that resist failure, classify the prompt type and apply a matching persuasion framework to reduce undesired outputs. This approach manages degrees of freedom and context-window economics for reliable behavior.

What is the best way to critique an existing LLM prompt draft?▼

The best way to critique an LLM prompt draft is using a dedicated critique mode to rewrite it against common failure modes. This evaluates prompt types, persuasion frameworks, and anti-pattern guidance to ensure safe behavior.

How do I classify LLM prompt types for different tasks?▼

LLM prompt types are classified into Discipline-enforcing, Guidance, Collaborative, and Reference categories. Tailoring behavior to these specific types allows you to apply deliberate persuasion frameworks and manage degrees of freedom appropriately.

How does context window budgeting work when creating sub-agent instructions?▼

Context window budgeting for sub-agent instructions works by selecting degrees of freedom based on task fragility. Low, medium, or high freedom directly manages the context-window economics to keep prompt behavior predictable.

When should I apply an ethics test to my prompt engineering workflow?▼

You should apply an ethics test during prompt engineering when finalizing design and critique workflows. This ensures safe, trustworthy prompts by providing explicit anti-pattern guidance before deploying hooks or sub-agent instructions.

Related Skills

Explore More Design & Creative Skills→
unfinishablemapunfinishablemap
3

expand-topic

Generate a publish-ready article draft from a topic prompt.

Community
Intermediate
daeha76daeha76
2

cc-dev-agent

Implement context engineering, spec-driven development, TDD, and sub-agent workflows for Claude Code.

Community
Advanced
Nuwanda04Nuwanda04

senior-prompt-engineer

Analyze prompts, evaluate RAG systems, and design agentic workflows.

Community
Advanced
zhuangwenhuizhuangwenhui
98

customize-agent:prompt-engineering

Craft effective prompts and commands for LLM interactions.

Community
Advanced