home / skills / technickai / ai-coding-config / writing-for-llms

writing-for-llms skill

safe

This skill applies comprehensive prompt engineering standards to craft effective LLM prompts, instructions, and Task tool prompts with clear goals.

npx playbooks add skill technickai/ai-coding-config --skill writing-for-llms

Review the files below or copy the command above to add this skill to your agents.

Files (1)

SKILL.md

1.1 KB

---
name: writing-for-llms
# prettier-ignore
description: Use when writing prompts, agent instructions, SKILL.md, commands, system prompts, Task tool prompts, prompt engineering, or LLM-to-LLM content
version: 1.0.0
category: meta
triggers:
  - "prompt"
  - "write prompt"
  - "system prompt"
  - "agent instructions"
  - "SKILL.md"
  - "prompt engineering"
  - "Task tool"
---

Apply the full prompt engineering standards from @rules/prompt-engineering.mdc

<key-principles>
- Show correct patterns only - never show anti-patterns, even labeled "wrong"
- State goals, not process - trust the executing model's capabilities
- Use XML tags for structure in complex prompts
- Clarity over brevity - every word that adds clarity is worth including
- Few-shot examples must follow identical structure
- Front-load critical information
- Consistent terminology throughout
</key-principles>

<quality-check>
Before finalizing:
- No anti-patterns shown anywhere
- All examples structurally consistent
- Goals clear, process not over-prescribed
- Terminology consistent
- Critical info front-loaded
</quality-check>

Overview

This skill helps you craft high-quality prompts, agent instructions, system prompts, Task tool prompts, and LLM-to-LLM content for reliable execution. It encodes strict prompt-engineering standards so outputs are clear, consistent, and safe for downstream models. Use it to standardize prompt assets across agents and tooling.

How this skill works

The skill inspects prompt drafts and enforces core principles: front-loading critical info, stating goals instead of prescribing process, and keeping terminology consistent. It restructures prompts into clear, machine-friendly formats, applies XML tagging for complex sections, and validates that few-shot examples are structurally identical. A final quality check confirms no anti-patterns, consistent examples, and preserved goals.

When to use it

Writing or revising system prompts and global agent instructions
Designing Task tool prompts and LLM-to-LLM handoffs
Creating few-shot examples for model conditioning
Standardizing prompt templates across a team or product
Auditing prompts for clarity, consistency, and safety

Best practices

Show only correct patterns; never include anti-patterns or labeled ‘wrong’ examples
State the goal or desired output format; do not micromanage the model's internal steps
Front-load the most critical constraints and success criteria
Use XML tags to structure complex prompts so sections are unambiguous
Ensure all few-shot examples share identical structure and terminology
Run a final quality check for anti-patterns, structural consistency, and clear goals

Example use cases

Convert a natural-language requirement into a concise system prompt with XML sections for context, task, and output schema
Audit an agent instruction set to remove anti-patterns and replace process prescriptions with outcome-focused goals
Create a set of few-shot examples that demonstrate the exact output structure for a parsing task
Standardize Task tool prompts so downstream agents receive consistent field names and tags
Validate LLM-to-LLM prompts to ensure handoffs include clear success criteria and no conflicting terminology

FAQ

What counts as an anti-pattern?

Any example or instruction that demonstrates incorrect behavior, prescribes internal chain-of-thought, or encourages shortcuts that reduce reliability.

Why use XML tags?

XML-like tags make complex prompts machine-readable and reduce ambiguity in section boundaries and expected output fields.