agents-md
0This skill enforces research-backed best practices for agent-facing documentation, ensuring quality by keeping it concise. It analyzes project structure to provide relevant instructions without unnecessary filler.
894460Updated Today
Documentation
Quick InstallView Source
$ npx skills add https://github.com/getsentry/skills --skill agents-mdOverview
About This Skill
agents-md enforces best practices for agent-facing documentation by creating and maintaining concise AGENTS.md and CLAUDE.md files. It analyzes your project structure to provide relevant, research-backed instructions tailored to agent capabilities, eliminating unnecessary filler. The skill helps teams document their projects effectively for AI agents like Claude Code and Cursor, ensuring documentation stays focused and actionable. It follows the Agent Skills open format and integrates seamlessly into Sentry's documentation workflows.
How to Use
Install using the command: `$ npx skills add https://github.com/getsentry/skills --skill agents-md`
Use Cases
Create concise documentation for agent commands
Maintain project-specific documentation styles
Analyze repository structure for relevant documentation
Reduce clutter in agent documentation
Tags
Setup & Installation
Quick Install
Terminal
$ npx skills add https://github.com/getsentry/skills --skill agents-mdAlternative Install (Git Clone)
git clone https://github.com/getsentry/skillsRequirements
- Claude Code or compatible AI agent
Quick Start Guide
1
Install the Skill
Copy the install command above and run it in your terminal.
2
Open Your AI Agent
Launch Claude Code, Cursor, or your preferred AI coding agent.
3
Try It Out
Use the prompt template or examples below to test the skill.
4
Customize
Adapt the skill to your specific use case and workflow.
Usage Examples
Prompt Template
Act as a documentation specialist for [COMPANY]'s AI agent systems. Review the following project structure and generate concise, actionable instructions for the agent to follow: [PROJECT_STRUCTURE]. Ensure instructions are research-backed, omit filler, and focus on critical context only.
Example Output
# Agent Instructions for Project: "RetailBot - E-commerce Assistant" ## Core Objectives - **Primary Goal**: Assist customers with product searches, order tracking, and returns. - **Secondary Goal**: Escalate complex issues to human agents when needed. ## Agent Workflow ### 1. Product Search - **Input**: User provides a product query (e.g., "wireless headphones under $100"). - **Action**: Query the product database using the `search_products` function with filters for price range and category. - **Output**: Return a ranked list of 5-10 relevant products with prices, ratings, and direct links. ### 2. Order Tracking - **Input**: User provides an order ID (e.g., "ORD-2023-45678"). - **Action**: Call `get_order_status` with the order ID. - **Output**: Return a summary of the order status (e.g., "Shipped on 10/15/2023 via FedEx") and estimated delivery date. ### 3. Return Initiation - **Input**: User requests a return for an order ID (e.g., "ORD-2023-45678"). - **Action**: Call `initiate_return` with the order ID and reason (e.g., "Defective product"). - **Output**: Return a return label link and instructions for packaging the item. ## Escalation Rules - Escalate to human agent if: - User is frustrated (detected via sentiment analysis). - Issue involves billing disputes or fraud. - Request requires manual intervention (e.g., custom refunds). ## Constraints - Never share internal system errors with the user. - Always confirm user intent before proceeding with actions (e.g., "You want to return ORD-2023-45678 for a defective product. Is that correct?").
Find the right skills for your stack
Take a free 3-minute scan and get personalized AI skill recommendations.
Take free scan