What is the CLEAR framework in prompt writing?

CLEAR is a mnemonic for Context, Limitations, Expectations, Actions, and Resources. It ensures every prompt provides complete background, explicit constraints, defined success criteria, step-by-step instructions, and necessary file references.

Why do subagents need specific context patterns?

Subagents in Claude Code have no access to your previous conversation history. They start fresh each time, so you must explicitly pass context through file-based state transfer or embedded info.

What are common prompt anti-patterns this skill avoids?

It prevents 'Context Amnesia' (assuming the agent remembers), 'Vague Expectations' (missing success criteria), and 'Unbounded Scope' (not defining limits), leading to more predictable AI behavior.

How does this skill help with multi-step workflows?

It provides standardized 'State Transfer Patterns' where Agent 1 saves results to a specific directory and Agent 2 is instructed to read from that path, maintaining continuity across tasks.

Prompt Engineering for Agents

Name: Prompt Engineering for Agents
Author: Danik911

byDanik911

0•

Productivity & Workflow

Streamlines AI subagent delegation and prompt construction using the CLEAR framework to ensure context-rich, actionable instructions.

The Prompt Writing skill provides a systematic approach to communicating with AI coding agents and sub-agents, solving the common problem of context amnesia in multi-step workflows. By utilizing the CLEAR framework (Context, Limitations, Expectations, Actions, Resources), it ensures that every delegated task contains the necessary background, constraints, and success criteria for high-quality execution. This skill is essential for developers building complex multi-agent systems where precision, state transfer, and explicit boundaries are critical for maintaining project integrity.

Key Features

01Implements the CLEAR framework for structured prompt creation

02Includes pre-built templates for delegation, research, and code review

03Standardizes agent output formats for better machine readability

04Provides state transfer patterns for seamless multi-agent workflows

05Offers a quality checklist to eliminate common prompting anti-patterns

060 GitHub stars

Use Cases

01Standardizing state management and data handoff between sequential AI tasks

02Delegating complex implementation tasks to Claude subagents with full context

03Designing robust system prompts and instructions for custom UI agents

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add danik911/thesis_project prompt-writing

For use in Claude.ai and ChatGPT

Download Skill