writing-prompts
Use when crafting prompts for LLMs to improve response quality, accuracy, consistency, and task performance
npx skills add majiayu000/claude-skill-registry --skill writing-prompts-darthlinuxer-agentic-skills-2 --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Instructs the agent to engage in prompt engineering for LLM tasks, focusing on structured design, iterative refinement, and model-specific optimization to improve response quality, accuracy, consistency, and task performance.
How it works
Adopts an iterative RED-GREEN-REFACTOR workflow:
- Define Success Criteria: specify measurable output requirements and formats.
- Establish Baseline (RED): test without optimization and document failures.
- Draft Initial Prompt (Minimal Fix): start simple to address observed issues.
- Apply Techniques (GREEN): layer prompt engineering methods until criteria are met.
- Test & Iterate: assess with diverse inputs and edge cases.
- Refactor: improve clarity, token efficiency, robustness without altering success criteria.
It emphasizes common prompting techniques (system prompts, role prompting, chain-of-thought, few-shot, XML/structured tags, output formatting, constraints & guards) and provides a comprehensive implementation sequence from defining tests to refactoring for efficiency.
When to use it
Use when creating, debugging, or improving prompts for LLM tasks—especially for production prompts requiring reliability, detailed constraints, and cross-provider compatibility.
What it can touch
The skill references tools like system prompts, role prompts, and structured outputs (JSON/XML) as core methods; it does not specify external software execution beyond designing prompts and testing strategies.
Caveats
Highlights the need for iterative testing, token-conscious design, and adherence to structured output formats. Includes guardrails to prevent non-conforming text and to handle unrecognizable inputs gracefully.
# Writing Prompts ## Overview Prompt engineering is the systematic process of crafting effective prompts for large language models (LLMs) to achieve better, more reliable, and more targeted responses. This skill provides proven techniques for maximizing LLM capabilities through structured prompt design, iterative refinement, and model-specific optimization. **Core principle:** Great prompts are specific, contextual, and tested. They guide the model's reasoning process while leaving appropriate degrees of freedom for the task at hand. ## When to Use Use this skill when: - Creating initial prompts for new tasks with LLMs - Improving existing prompts that produce inconsistent or suboptimal results - Working with complex tasks requiring step-by-step reasoning - Needing LLMs to adopt specific roles or behaviors - Developing prompts for production applications - Debugging prompt-related issues in LLM interactions - Optimizing token usage while maintaining output quality - Adapting prompts across different LLM providers (Anthropic, GPT, Gemini, etc.) ## Core Pattern Follow this iterative RED-GREEN-REFACTOR process for prompt development: 1. **Define Success Criteria** (Write the Test): Cl
- Overview
- When to Use
- Core Pattern
- Fundamental Principles
- 1. Be Specific and Direct
- 2. Provide Sufficient Context
- 3. Match Freedom to Task Fragility
- 4. Show, Don't Just Tell
- 5. Guide the Reasoning Process
- Quick Reference
- Implementation
- Step 1: Establish Success Criteria (Write Your Test)
- Step 2: Test Baseline Behavior (RED)
- Step 3: Use System Prompts for Persistent Context
What does the writing-prompts skill do?
Use when crafting prompts for LLMs to improve response quality, accuracy, consistency, and task performance
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill writing-prompts-darthlinuxer-agentic-skills-2 --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.