<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/5-concepts-to-know-for-coding-with-ai-agents-vskpd1jto" -->

---
title: 5 Concepts to Know for Coding with AI Agents | daily.dev
description: Coding with AI agents effectively comes down to five concepts: context window management, planning, verification loops, rules files, and tools. A Claude...
canonical: https://daily.dev/posts/5-concepts-to-know-for-coding-with-ai-agents-vskpd1jto
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: 5 Concepts to Know for Coding with AI Agents | daily.dev
og:description: Coding with AI agents effectively comes down to five concepts: context window management, planning, verification loops, rules files, and tools. A Claude...
og:url: https://daily.dev/posts/5-concepts-to-know-for-coding-with-ai-agents-vskpd1jto
og:image: https://api.daily.dev/og/posts/VSkPD1jto.png
og:image:alt: 5 Concepts to Know for Coding with AI Agents
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# 5 Concepts to Know for Coding with AI Agents

**[Tech With Tim](https://daily.dev/sources/TechWithTim)** · 19 min read · 6 upvotes · 0 comments

## Summary

Coding with AI agents effectively comes down to five concepts: context window management, planning, verification loops, rules files, and tools. A Claude session can start at 71,000 tokens of context and balloon to over 180,000 tokens after a single prompt like 'read every file in my project.' Clearing unused MCP tools and scoping them per-project via an .mcp.json file reduces token bloat and agent confusion. Rules files (claude.md, agents.md, cursor rules) persist project-specific instructions like 'never use floats for money' across sessions. Verification loops, giving the agent a script to run tests and compilation, let it catch its own mistakes before declaring work done, and sub-agents can parallelize front-end, back-end, and API work simultaneously.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://www.youtube.com/watch?v=DfAOBID70Q0>

## Questions this post answers

### How can I stop Claude Code's context window from filling up too fast

Inspect your context before running anything using the slash context command, which shows a token breakdown of MCP tools, system prompts, skills, and server instructions. Disable unused MCP tools or scope them to a specific project instead of loading them globally. A single prompt like 'read every file in my project' can grow a 71,000-token session to over 180,000 tokens; use slash clear to reset it.

_Anyone tuning agent context limits can track workflow tips like this on daily.dev._

### How do I scope MCP servers to a single project instead of loading them globally in Claude Code

Create a .mcp.json file in the project root specifying only the MCP servers needed for that session, instead of connecting them globally. Global MCP connections can add 20,000 to 30,000 tokens to every session and cause the agent to pick the wrong tool among too many options. You can also load Claude with a specific config file, such as a separate session mcp.json, to swap tool sets per use case.

_Developers wiring up MCP servers per repo can follow setup patterns like this on daily.dev._

### What should I put in a claude.md or agents.md rules file

Include the app name and description, commands to run and reset the app, a required verification script to run before declaring work done, and project-specific rules such as 'money is stored as an integer, never do money math with floats' or 'never add an existing test just to make it pass.' Committing this file to the git repo lets every agent session and teammate follow the same guidance automatically.

_Teams standardizing agent behavior across a codebase can compare rules-file setups on daily.dev._

## Similar posts on daily.dev

- [Five Things Your Coding AI Agent Wishes You Understood](https://daily.dev/posts/five-things-your-coding-ai-agent-wishes-you-understood-ly3dv4pqh) · The Miners · 0 upvotes · 0 comments
- [How to Manage Context Files in Your Codebase and Get Better Output From AI Coding Agents](https://daily.dev/posts/how-to-manage-context-files-in-your-codebase-and-get-better-output-from-ai-coding-agents-jmjpmaeus) · freeCodeCamp · 2 upvotes · 1 comments

---

Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#llm](https://daily.dev/tags/llm), [#mcp](https://daily.dev/tags/mcp), [#claude-code](https://daily.dev/tags/claude-code), [#prompt-engineering](https://daily.dev/tags/prompt-engineering)

[View this post on daily.dev](https://daily.dev/posts/5-concepts-to-know-for-coding-with-ai-agents-vskpd1jto)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"5 Concepts to Know for Coding with AI Agents","url":"https://daily.dev/posts/5-concepts-to-know-for-coding-with-ai-agents-vskpd1jto","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/5-concepts-to-know-for-coding-with-ai-agents-vskpd1jto"},"datePublished":"2026-11-17T13:13:07.811Z","dateModified":"2026-11-17T13:13:33.778Z","description":"Coding with AI agents effectively comes down to five concepts: context window management, planning, verification loops, rules files, and tools. A Claude...","image":"https://i.ytimg.com/vi/DfAOBID70Q0/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/DfAOBID70Q0/sddefault.jpg","isAccessibleForFree":true,"articleSection":"Tech With Tim","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Tech With Tim","logo":"https://media.daily.dev/image/upload/s--6vVIDfU0--/f_auto/v1711726146/logos/TechWithTim","url":"https://daily.dev/sources/TechWithTim"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/5-concepts-to-know-for-coding-with-ai-agents-vskpd1jto","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":6},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,llm,mcp,claude-code,prompt-engineering","timeRequired":"PT19M","video":{"@type":"VideoObject","name":"5 Concepts to Know for Coding with AI Agents","description":"Coding with AI agents effectively comes down to five concepts: context window management, planning, verification loops, rules files, and tools. A Claude...","thumbnailUrl":"https://i.ytimg.com/vi/DfAOBID70Q0/sddefault.jpg","uploadDate":"2026-11-17T13:13:07.811Z","duration":"PT19M","url":"https://api.daily.dev/r/VSkPD1jto","embedUrl":"https://www.youtube.com/embed/DfAOBID70Q0"}}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Tech With Tim","item":"https://daily.dev/sources/TechWithTim"},{"@type":"ListItem","position":3,"name":"5 Concepts to Know for Coding with AI Agents"}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/5-concepts-to-know-for-coding-with-ai-agents-vskpd1jto#faq","mainEntity":[{"@type":"Question","name":"How can I stop Claude Code's context window from filling up too fast","acceptedAnswer":{"@type":"Answer","text":"Inspect your context before running anything using the slash context command, which shows a token breakdown of MCP tools, system prompts, skills, and server instructions. Disable unused MCP tools or scope them to a specific project instead of loading them globally. A single prompt like 'read every file in my project' can grow a 71,000-token session to over 180,000 tokens; use slash clear to reset it. Anyone tuning agent context limits can track workflow tips like this on daily.dev."}},{"@type":"Question","name":"How do I scope MCP servers to a single project instead of loading them globally in Claude Code","acceptedAnswer":{"@type":"Answer","text":"Create a .mcp.json file in the project root specifying only the MCP servers needed for that session, instead of connecting them globally. Global MCP connections can add 20,000 to 30,000 tokens to every session and cause the agent to pick the wrong tool among too many options. You can also load Claude with a specific config file, such as a separate session mcp.json, to swap tool sets per use case. Developers wiring up MCP servers per repo can follow setup patterns like this on daily.dev."}},{"@type":"Question","name":"What should I put in a claude.md or agents.md rules file","acceptedAnswer":{"@type":"Answer","text":"Include the app name and description, commands to run and reset the app, a required verification script to run before declaring work done, and project-specific rules such as 'money is stored as an integer, never do money math with floats' or 'never add an existing test just to make it pass.' Committing this file to the git repo lets every agent session and teammate follow the same guidance automatically. Teams standardizing agent behavior across a codebase can compare rules-file setups on daily.dev."}}]}
```

