mcp AGENTS.md

A set of instructions for AI coding agents working on Microsoft’s MCP project. It defines coding rules, command patterns, testing requirements, and Azure resource conventions.

In plain words
What is it for?
Adding or changing C# Azure commands, naming resources and options, creating Bicep test resources, writing post-deployment scripts, and running the required tests. Azure is Microsoft’s cloud platform, and MCP connects assistants to tools.
Why use it?
It keeps changes consistent with the repository’s expectations and reminds contributors about builds, live tests, deployment files, and error handling.

Instructions file for CodexOpenCode

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/microsoft/mcp/agents-md
Clone the repo
git clone --depth 1 https://github.com/microsoft/mcp

Made for: Codex, OpenCode.

Per session 7,571 This file is loaded in full into every session.
When invoked 7,571 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.07571 $0.07571
Opus 5 $0.03785 $0.03785
Sonnet 5 $0.01514 $0.01514
Haiku 4.5 $0.00757 $0.00757

Measured 2d ago against content hash f5328a2654a4, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mcp AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

AGENTS.md · 786 lines

How it starts

The opening of the file, as written. The whole thing — 786 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

Do

  • Use primary constructors for all C# classes
  • Use System.Text.Json
  • Make command classes sealed unless designed for inheritance
  • Make members static when possible for AOT compatibility
  • Put each class and interface in separate files
  • Use subscription parameter (never subscriptionId) - supports both IDs and names
  • Use resourceGroup (not resourceGroupName)
  • Use singular nouns for resource names (e.g., server, not serverName)
  • Run dotnet build after making changes
  • Follow the {Resource}{Operation}Command naming pattern
  • Use the two-generic pattern SubscriptionCommand<TOptions, TResult> for new commands
  • Use [Option] attributes on flat POCO option classes for option declaration
  • Use ISubscriptionOption interface on options classes that need subscription/tenant
  • Always call HandleException(context, ex) in catch blocks
  • Include live tests for all commands that interact with Azure resources
  • Create Bicep templates for Azure service commands (test-resources.bicep)
  • Include post-deployment scripts (test-resources-post.ps1)
  • Record all live tests according to the guidelines in /docs/recorded-tests.md
  • Submit one tool per pull request
  • Use BaseAzureResourceService for Resource Graph queries when possible
  • Register all response models in JSON serialization context for AOT safety
  • Register all commands in the appropriate Setup.cs file
  • Use concatenated lowercase for command group names (no dashes)
  • Prefer file-scoped changes over project-wide modifications when possible
  • Always review your own code for consistency, maintainability, and testability
  • Always ask for clarifications if the request is ambiguous or lacks sufficient context
  • Write transport-agnostic commands that work in both stdio and HTTP modes
  • Keep commands stateless and thread-safe for multi-user remote scenarios
  • Test commands with different RBAC permissions for OBO scenarios

Don't

  • Use subscriptionId parameter name
  • Add unnecessary "-name" suffixes (use --account vs --account-name)
  • Use the old one-generic RegisterOptions/BindOptions pattern for new commands (see /docs/option-conversion.md)
  • Use OptionDefinitions static classes for new commands (use [Option] attributes instead)
  • Use options class inheritance hierarchies (use flat POCOs with interface constraints)
  • Skip live tests, live test infrastructure, or test recordings for Azure service commands
  • Redefine base class properties in Options classes
  • Leave commands unregistered
  • Skip error handling or comprehensive tests
  • Use dashes in command group names (use concatenated lowercase)
  • Make project-wide changes when file-scoped changes suffice
  • Check transport type in commands (stdio vs HTTP)
  • Store per-request state in command instance fields
  • Access HttpContext directly from commands
  • Make transport-specific decisions in command logic
  • Assume single-user scenarios when implementing services
  • Use git stash in linked worktrees; the stash stack is shared across worktrees and can collide during concurrent work

Read the full file on GitHub · 786 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 786 lines · 7,571 tokens per session scan A f5328a2654a4

Subscribe to this mod's changes

mcp AGENTS.md is an instructions file published in the GitHub repository microsoft/mcp (3,625 stars, last pushed 3d ago), licensed MIT. It adds 7,571 tokens to every session, about $0.0379 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other instructions, from other repositories

codex AGENTS.md

AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.

openai/codex · 5,182 tokens

buildNext

Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).

microsoft/vscode · 6,785 tokens

next.js AGENTS.md

Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.

vercel/next.js · 7,296 tokens

vscode oss-third-party-notices.instructions.md

Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).

microsoft/vscode · 5,001 tokens

spec-kit AGENTS.md

Instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.

github/spec-kit · 7,040 tokens

langchain AGENTS.md

Instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.

langchain-ai/langchain · 4,345 tokens