robots.txt
See which AI crawlers other sites block →User-agent: *
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: GPTBot
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: OAI-SearchBot
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: ChatGPT-User
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: ClaudeBot
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: Claude-Web
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: anthropic-ai
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: Claude-SearchBot
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: PerplexityBot
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: Perplexity-User
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: Google-Extended
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: CCBot
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: Applebot-Extended
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: Amazonbot
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: Meta-ExternalAgent
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: Bytespider
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
User-agent: cohere-ai
Allow: /
Disallow: /admin/
Disallow: /api/
Disallow: /_next/
Disallow: /static/
Disallow: /cursor*
Disallow: /reddit*
Disallow: /windsurf*
Disallow: /x*
Disallow: /li*
Disallow: /vibe-coders-hq*
Disallow: /*?dub_id=*
Disallow: /*?utm_*
Disallow: /*?ref=*
Disallow: /c$
Disallow: /health$
Disallow: /security.asc$
Disallow: /hub/
Disallow: /hub/*
Disallow: /waitlist
Disallow: /blog/headers/*
Disallow: /blog/*/opengraph-image
Disallow: /logos/*
Disallow: /brand.txt
Disallow: /dashboard
Content-Signal: search=yes, ai-input=yes, ai-train=yes
Sitemap: https://www.braingrid.ai/sitemap.xml
Host: https://www.braingrid.ai
llms.txt
Compare with other sites' llms.txt files →# BrainGrid
> BrainGrid is the AI Product Planner that helps builders turn rough ideas into engineering-grade tasks their AI coding tools can build right the first time. Plan, build, verify, repeat.
BrainGrid works with any tool that supports MCP (Model Context Protocol) — including Claude Code, Cursor, Windsurf, OpenAI Codex, Google Gemini, and Lovable. This file curates the site's key content for LLMs and agents. Most content pages also expose raw markdown at the same URL with a `.md` suffix.
## Core Pages
- [Home](https://www.braingrid.ai/index.md): BrainGrid overview — plan, build, verify, repeat.
- [Agentic Engineering](https://www.braingrid.ai/agentic-engineering.md): The workflow that makes AI coding work — The BrainGrid Loop: Frame, Spec, Scope, Load, Verify.
- [Spec-Driven Development](https://www.braingrid.ai/spec-driven-development.md): Plan once, build right — write the spec first, let your coding agent build against it.
- [AI PRD Generator](https://www.braingrid.ai/ai-prd-generator.md): From rough idea to build-ready requirements with acceptance criteria and tasks.
- [Vibe Coding Tools](https://www.braingrid.ai/vibe-coding-tools.md): Honest guide to Lovable, Bolt, Replit, v0, Base44, Cursor, and Claude Code.
- [AI Coding Agents](https://www.braingrid.ai/ai-coding-agents.md): How coding agents work and how BrainGrid loads every one of them.
- [The BrainGrid Loop](https://www.braingrid.ai/loop.md): The named method — Frame, Spec, Scope, Load, Verify.
- [Context Engineering](https://www.braingrid.ai/context-engineering.md): Feed your agent facts, not guesses — a spec is engineered context.
- [AI for Product Managers](https://www.braingrid.ai/ai-for-product-managers.md): Turn roadmap items into requirements, then into pull requests.
- [BrainGrid vs Spec Kit](https://www.braingrid.ai/braingrid-vs-spec-kit.md): Honest comparison of two takes on spec-driven development.
- [Pricing](https://www.braingrid.ai/pricing.md): Plans for solo builders and teams from $29/mo per seat. No free trial; cancel anytime.
- [Community](https://www.braingrid.ai/community.md): Free Slack community for AI builders.
- [Blog](https://www.braingrid.ai/blog): AI coding, product planning, and developer productivity.
- [Brand Guidelines](https://www.braingrid.ai/brand.md): Logos, colors, typography, and voice.
- [Talk to a Human](https://www.braingrid.ai/talk-to-a-human.md): Contact the BrainGrid team.
## Resources
- [Documentation](https://www.braingrid.ai/docs): Product docs and guides. Every page has a markdown twin at its URL + `.md` (e.g. https://www.braingrid.ai/docs/quickstart.md).
- [RSS Feed](https://www.braingrid.ai/feed.xml): Latest blog posts.
- [Sitemap](https://www.braingrid.ai/sitemap.xml): Full URL index.
- [llms-full.txt](https://www.braingrid.ai/llms-full.txt): Full markdown of the docs, all blog posts and legal docs.
## Blog
- [BDD, Explained: Given-When-Then Is Just Acceptance Criteria Your Agent Can Read](https://www.braingrid.ai/blog/bdd-given-when-then.md): Behavior-driven development's Given-When-Then format was built for a meeting. Its real payoff arrives now that the reader is an AI agent. Here is why.
- [Agent Evals for Builders: Your Acceptance Criteria Are the Rubric](https://www.braingrid.ai/blog/agent-evals-for-builders.md): Agent evals sound like ML infrastructure. For builders shipping features, an eval is just a rubric the agent's output is checked against. Here's the version you actually need.
- [What Counts as Done? A Founder's Guide to the Definition of Done](https://www.braingrid.ai/blog/what-counts-as-done-founders-guide.md): A plain-language definition of done for founders who ship with AI agents. Why 'it looks finished' is the trap, and how to decide done before you build.
- [What Is a Database Migration? The One Change Your AI Agent Makes That Can Lose Real Data](https://www.braingrid.ai/blog/what-is-a-database-migration.md): What is a database migration, in plain language for AI builders: the one change your coding agent makes that can drop a column or lose rows for real, and the criterion that keeps it safe.
- [What Is a Modular Monolith? The App Structure That Keeps Your AI Agent From Hitting the Complexity Wall](https://www.braingrid.ai/blog/what-is-a-modular-monolith.md): What is a modular monolith, in plain language for AI builders: one app with clear internal boundaries, and why it is the structure that stops your agent from fixing one thing and breaking ten as your codebase grows.
- [How Much Time Should You Spend Planning Before Coding With AI?](https://www.braingrid.ai/blog/how-much-time-planning-before-coding-ai.md): AI made building cheap, so planning should matter less. It matters more. Here's the ratio real builders land on, and why unplanned time comes back with interest.
- [The Best Spec-Driven Development Tools in 2026 (Tested & Compared)](https://www.braingrid.ai/blog/spec-driven-development-tools.md): A field guide to spec-driven development tools: Spec Kit, OpenSpec, Kiro, BMAD, and BrainGrid. What each one actually does, and how to pick the right one.
- [How to Verify AI-Generated Code (Before It Breaks Production)](https://www.braingrid.ai/blog/how-to-verify-ai-generated-code.md): A practical guide to verifying AI-generated code: the checks that catch what the agent quietly broke, and why 'it runs' is not the same as 'it's done.'
- [EARS Notation, Explained: The Requirements Syntax Behind Every AI Spec](https://www.braingrid.ai/blog/ears-notation.md): EARS notation is the Easy Approach to Requirements Syntax: a five-pattern template for writing requirements an AI agent can actually verify. How it works.
- [What Is a CLAUDE.md File? The One Anthropic's Own Docs Tell You to Keep Short](https://www.braingrid.ai/blog/what-is-a-claude-md-file.md): What a CLAUDE.md file is, what goes inside it, whether to commit it, and the one thing it was never meant to be: your product's record.
- [What Is AI-Native Development? (And Why "Disposable Software" Is Only Half True)](https://www.braingrid.ai/blog/ai-native-development.md): AI-native development means building software assuming the agent writes most of it. An honest definition, and why disposable software is only half true.
- [What Is an Agent Loop? The Reason-Act-Observe Cycle, Explained](https://www.braingrid.ai/blog/what-is-an-agent-loop.md): An agent loop is the reason, act, observe cycle that turns a chatbot into an agent. What each step does, when it stops, and the one input it can't supply itself.
- [What Is a System Prompt? Where Your Durable Instructions Actually Belong](https://www.braingrid.ai/blog/what-is-a-system-prompt.md): A system prompt sets how your AI agent behaves. Learn what it is, how it differs from a user prompt, what belongs in it, and where your product's real requirements should live instead.
- [The Illusion of Completeness: Why Your AI Demo Feels Done and Isn't](https://www.braingrid.ai/blog/illusion-of-completeness-ai-demo.md): Your AI-built demo looks finished, so why are you scared to ship it? The gap between a prototype that runs and a product that works, and how to close it.
- [Architecture Decision Records for AI Coding Agents: Where the Why Lives](https://www.braingrid.ai/blog/architecture-decision-records-for-ai-coding-agents.md): An architecture decision record captures why you chose an approach. For AI coding agents that forget between sessions, the ADR is the one artifact that survives.
- [The Slot Machine Trap: Why AI Coding Burns You Out (and Ships Nothing)](https://www.braingrid.ai/blog/slot-machine-trap-ai-coding-burnout.md): Vibe coding problems get blamed on bad code. The real one is a loop that never converges. Why AI coding burns you out, and how to stop pulling the lever.
- [UAT for Solo Builders: Why Your Own Test Accounts Pass Every Check](https://www.braingrid.ai/blog/uat-for-solo-builders.md): UAT assumes a tester who isn't the builder. When you're the whole QA department, your own accounts pass every check. Here's the solo-builder version that works.
- [Requirements Management for AI Builders (Without the Enterprise Baggage)](https://www.braingrid.ai/blog/requirements-management-for-ai-builders.md): Requirements management used to mean a heavyweight enterprise tool. For AI builders, it's the plan your agent builds against. Here's the version that fits.
- [Vibe Coding vs Agentic Coding: Which One Actually Ships?](https://www.braingrid.ai/blog/vibe-coding-vs-agentic-coding.md): Vibe coding vs agentic coding is not a fight about autonomy or tooling. It is about who owns the definition of done. Here is where the line really falls.
- [What Is a Merge Conflict? The Non-Engineer's Fix When Your AI Agent Gets Stuck](https://www.braingrid.ai/blog/what-is-a-merge-conflict.md): Your AI agent hit a merge conflict and stopped. What a merge conflict is in plain English, why agents cause more of them, and how to fix one without editing code.
- [Why We Made Our Agent Ask Questions Before It Builds](https://www.braingrid.ai/blog/why-we-made-our-agent-ask-questions.md): The best AI coding agent isn't the one that builds fastest from your prompt. It's the one that asks the questions you didn't think to answer, before it builds.
- [Agent Orchestration for Builders: When One Agent Isn't Enough](https://www.braingrid.ai/blog/agent-orchestration-for-builders.md): Agent orchestration coordinates multiple AI agents toward one goal. What it is, the four patterns that matter, and the input no orchestrator can generate for you.
- [Why Reviewing AI Code Costs More Than Writing It](https://www.braingrid.ai/blog/why-reviewing-ai-code-costs-more-than-writing-it.md): Reviewing AI generated code often takes longer than writing it yourself. Here is why the review tax is real, and how to move that time earlier instead of paying it twice.
- [What Is an Agent Flywheel? Why Most Spin Fast and Compound Nothing](https://www.braingrid.ai/blog/agent-flywheel.md): An agent flywheel is a loop where every pass makes the next one easier. Here's what makes one actually compound, and why most just spin fast and accumulate nothing.
- [The Three Artifacts That Actually Matter: Requirements, Acceptance Tests, Code](https://www.braingrid.ai/blog/three-artifacts-requirements-acceptance-tests-code.md): Requirements, acceptance tests, and code are the only three artifacts that carry signal in an AI build. Why acceptance tests are the one most builders skip.
- [The Real Cost of a Vague Prompt (In Tokens, Credits, and Afternoons)](https://www.braingrid.ai/blog/the-real-cost-of-a-vague-prompt.md): The real cost of AI coding is not the token price. A vague prompt gets re-run, reviewed, and reworked. Here is where the money goes, and how a plan cuts it.
- [Claude Code vs Cursor vs Codex vs Gemini: The 2026 Agent Comparison](https://www.braingrid.ai/blog/claude-code-vs-cursor-vs-codex-vs-gemini.md): An honest comparison of the four major AI coding agents in 2026, and the uncomfortable finding underneath the benchmarks: the agent you pick matters less than what you hand it.
- [Skills, Hooks, and Subagents: Assembling Your First Agent Loop in Claude Code](https://www.braingrid.ai/blog/skills-hooks-subagents-first-agent-loop.md): What Claude Code hooks, skills, and subagents do, how they assemble into an agent loop, and the one thing all three cannot give the loop: a definition of done.
- [Webhook vs API: What Your Agent Wired Up When It Added a Webhook](https://www.braingrid.ai/blog/webhook-vs-api.md): Webhook vs API in plain language for AI builders: an API is a door you knock on to pull data, a webhook is one another service knocks on to push it, plus the check your agent skipped.
- [What Is an API Gateway? The Front Door Your AI Agent Built for a Backend You Didn't Design](https://www.braingrid.ai/blog/what-is-an-api-gateway.md): What is an API gateway, in plain language for AI builders: the single front door your coding agent wired up, what it enforces, and why those rules are criteria you specify.
- [AGENTS.md, Explained: The README Your Coding Agent Actually Reads](https://www.braingrid.ai/blog/agents-md-explained.md): What is AGENTS.md, in plain language: the standard file that tells your AI coding agent how your project works, what belongs in it, what does not, and where it quietly falls short.
- [What Is an API Contract? The Agreement That Keeps Your AI Agent From Breaking Its Own Codebase](https://www.braingrid.ai/blog/what-is-an-api-contract.md): What is an API contract, in plain language for AI builders: the written agreement between two parts of your app, and why it is the boundary that keeps your agent from breaking things as your codebase grows.
- [The 8 Questions to Answer Before Your AI Agent Writes a Line of Code](https://www.braingrid.ai/blog/8-questions-before-your-agent-writes-code.md): What to plan before building an app with AI: the eight questions that decide whether your agent ships a product or a mess you can't trust.
- [Context Rot: Why Your AI Agent Gets Dumber the Longer It Runs](https://www.braingrid.ai/blog/context-rot.md): Context rot is why AI coding agents forget files and break working code as sessions grow. What causes it, why bigger windows don't fix it, and how to stop it.
- [Agentic Engineering vs Vibe Coding: The Line Is the Plan, Not the Vocabulary](https://www.braingrid.ai/blog/agentic-engineering-vs-vibe-coding.md): Agentic engineering vs vibe coding isn't about the tool you use. The real difference is whether a reviewable plan and a gate existed before the agent ran.
- [The Best Code Review Skill Is the One You Never Have to Run](https://www.braingrid.ai/blog/best-code-review-skill-review-against-a-standard.md): Everyone is installing code review skills for their AI agent. The best code review skill is a standard you set before the build, not a checker you run after.
- [What Is an API Endpoint? The Door Your Agent Built Into Your App](https://www.braingrid.ai/blog/what-is-an-api-endpoint.md): What is an API endpoint, in plain language for AI builders: the public door your coding agent added, who is allowed through it, and why that is a criterion you specify.
- [What Is Row-Level Security? The Auth Hole in Your Vibe-Coded App That Isn't a Code Bug](https://www.braingrid.ai/blog/what-is-row-level-security-for-vibe-coders.md): What is row level security, in plain language for vibe coders: the Supabase auth hole that leaks other users' data is a missing access criterion, not a code bug.
- [Vibe Coding Security: The Holes Aren't in the Code, They're in the Spec](https://www.braingrid.ai/blog/vibe-coding-security.md): Vibe coding security fails because the common holes, missing auth, no row-level security, hardcoded secrets, are unspecified acceptance criteria, not scanner findings.
- [How to Vibe Code the Right Way: Plan First, Then Let It Rip](https://www.braingrid.ai/blog/how-to-vibe-code-the-right-way.md): How to vibe code without the three-month mess: plan the feature first, then let the agent build. A practical method that keeps the speed and loses the debt.
- [The Fallback Trap: When Defensive AI Code Hides a Broken App](https://www.braingrid.ai/blog/the-fallback-trap.md): Defensive programming should make software safer. When an AI agent does it, fallbacks often hide a broken state so the app looks healthy while doing the wrong thing.
- [OpenSpec vs Spec Kit vs BrainGrid: The Honest Decision Map](https://www.braingrid.ai/blog/openspec-vs-spec-kit-vs-braingrid.md): OpenSpec vs Spec Kit vs BrainGrid: a straight decision map for choosing a spec-driven development tool, and the one question that actually decides it.
- [Human in the Loop vs On the Loop: Where You Belong When Agents Build](https://www.braingrid.ai/blog/human-in-the-loop-vs-on-the-loop.md): Human in the loop vs on the loop, explained for AI builders: what each posture means, when to use which, and the one thing that lets you move from in to on.
- [Software Development Methodologies in the AI Era: Where Spec-Driven Fits](https://www.braingrid.ai/blog/software-development-methodologies-ai-era.md): A field guide to software development methodologies, from Waterfall and Agile to spec-driven development, and where spec-driven fits in the AI era.
- [OpenSpec vs Spec Kit vs SuperPowers: Stop Shopping for the Right SDD Tool](https://www.braingrid.ai/blog/stop-shopping-for-spec-driven-development-tools.md): Confused by OpenSpec, Spec Kit, and SuperPowers? The spec-driven development tool matters far less than the shared practice underneath all three.
- [Definition of Done for AI Builders: The Quality Bar Every Agent Build Must Clear](https://www.braingrid.ai/blog/definition-of-done-for-ai-builders.md): A definition of done is the standing quality bar every feature must clear. Why AI builders need one, how it differs from acceptance criteria, and what goes in it.
- [GitHub Spec Kit Tutorial: Spec-Driven Development in an Existing Project](https://www.braingrid.ai/blog/github-spec-kit-tutorial-existing-project.md): A practical GitHub Spec Kit tutorial for an existing codebase: the six commands, a real constitution, and the point where Spec Kit stops and you take over.
- [Is Vibe Coding Good? An Honest Answer (Great Start, Bad Finish)](https://www.braingrid.ai/blog/is-vibe-coding-good.md): Is vibe coding good? Yes, for getting started. The honest answer is that it is a great start and a bad finish, and here is exactly where the line falls.
- [Every Software Factory Needs a Planning Floor](https://www.braingrid.ai/blog/every-software-factory-needs-a-planning-floor.md): A software factory automates the line that turns intent into shipped code. The part nobody automates is where intent gets made precise. That floor decides everything downstream.
- [The Four Pillars of Context Engineering: Write, Select, Compress, Isolate](https://www.braingrid.ai/blog/four-pillars-of-context-engineering.md): The four pillars of context engineering, Write, Select, Compress, Isolate, explained for builders who ship with coding agents, plus the one pillar everyone skips.
- [The AI Coding Workflow That Ships One Feature: Idea to Merged PR](https://www.braingrid.ai/blog/ai-coding-workflow-idea-to-merged-pr.md): A real AI coding workflow that takes one feature from idea to a merged pull request, and the one piece that makes it work when you can't read the code.
- [Why Your AI Agent Keeps Breaking Things That Worked](https://www.braingrid.ai/blog/why-ai-keeps-breaking-things-that-worked.md): Your AI coding agent keeps breaking features that already worked. Here is why it happens, and the one habit that stops the regressions for good.
- [The Verification Gap: Code Got Cheap, Trust Didn't](https://www.braingrid.ai/blog/verification-gap.md): The verification gap is the distance between code an AI can generate and code you can actually trust. Here is why it is widening and how to close it.
- [Kiro vs BrainGrid: Two Takes on Spec-Driven Development](https://www.braingrid.ai/blog/kiro-vs-braingrid-spec-driven-development.md): Kiro spec-driven development bakes specs into an AI IDE. BrainGrid puts them upstream of any agent. An honest comparison of where each one fits.
- [Claude Code Memory: The State File Pattern (and Where It Falls Short)](https://www.braingrid.ai/blog/claude-code-memory-state-file-pattern.md): How Claude Code memory actually works, why auto memory goes stale, and the state file pattern that gives your coding agent durable context across sessions.
- [AI Task Management for Builders Who Ship With Agents](https://www.braingrid.ai/blog/ai-task-management-for-builders-who-ship-with-agents.md): AI task management isn't a smarter to-do list. For builders shipping with coding agents, it's the work queue that turns a plan into tasks an agent can verify.
- [Structured Prompting Is Just a Spec You Haven't Written Yet](https://www.braingrid.ai/blog/structured-prompting-is-just-a-spec.md): Structured prompting organizes a prompt into role, context, task, constraints