Guides
The questions a developer asks before they know what Autoplans is: how MCP works, which servers to connect, what to put in CLAUDE.md and AGENTS.md, and why an agent keeps following instructions that stopped being true.
The Model Context Protocol, explained for people who have to wire it up
What an MCP server is, how JSON-RPC tool calls work, the difference between stdio and Streamable HTTP transport, and what a client actually sends the model.
MCP servers worth connecting to a coding agent
The servers that do real work in a Claude Code or Cursor setup — repositories, databases, browsers, files, and the project backlog — with what each one exposes.
Why a coding agent needs a backlog, not a prompt
Long-horizon tasks, multi-agent sessions and sprint planning all fail the same way when the plan lives in a context window. What to give the agent instead.
AGENTS.md: what to put in it, how to structure it, and when to use a skill instead
A working template for the instruction file OpenCode, Codex and other agents read first, the sections that earn their place, and the ones that rot.
.cursorrules and .cursor/rules: examples that hold up
How Cursor reads its rules files, a template for a Next.js project, what to generate versus write by hand, and how to keep the rules from going stale.
GitHub Copilot custom instructions: configuring copilot-instructions.md
Where the file goes, what Copilot does with it, an example for a TypeScript repository, prompt files, and how to generate one from the project itself.
CLAUDE.md: templates and the project context Claude Code actually uses
How Claude Code loads CLAUDE.md, a template for a Next.js project, what belongs in it, and how to point it at the plan instead of copying the plan into it.
Preventing AI context rot: why agents follow outdated instructions
Instruction files describe a moment and the project moves on. How to tell what rots, how to keep the durable part small, and where to keep the part that changes.