Reference guide for writing agent prompts and Warden skills.
npx skills add https://github.com/getsentry/warden --skill agent-prompt-getsentry
Please help me install this Agent Skill. Skill: agent-prompt Source: https://github.com/getsentry/warden/tree/main/.agents/skills/agent-prompt Command: npx skills add https://github.com/getsentry/warden --skill agent-prompt-getsentry
This Skill provides a comprehensive guide to crafting effective prompts for AI agents and developing robust Warden skills, ensuring clear communication and optimal performance.
Use the agent-prompt skill to learn how to write a new skill for Warden.
High-intent search queries and answers about installing and using this skill.
Best practices for writing effective agent prompts involve following foundational rules for clarity, establishing a well-defined prompt architecture, specifying output formats, and avoiding common anti-patterns to ensure optimal performance.
To structure a Warden skill for Claude 4.x models, you must define the skill's core components, apply model-specific guidance, and ensure the task is well-defined with necessary exclusions included for clear communication.
Common anti-patterns in prompt engineering include poorly defined tasks, lack of structural clarity, missing necessary exclusions, and ignoring model-specific guidance, all of which degrade agent performance and communication.
Specifying output formats is required when creating agent skills to ensure clear communication and optimal performance. Defining the expected output structure helps maintain consistency and accuracy in agent responses.
You should consult a reference guide for prompt architecture when you need to craft a new Warden skill, analyze code for security vulnerabilities, or ensure your prompt wording is clear, concise, and free of anti-patterns.
mwayDefine analysis scope, objectives, constraints, and success criteria with evidence and risk assessment.
grichaAnalyze local code changes with Warden to generate structured findings.
tonone-aiOrchestrate multi-agent engineering workflows with lead and specialist roles.
faionfaionGuide LLM integration, RAG pipelines, agent development, and model evaluation.
