Great tips, but check the date, they get stale quickly.

• This One Prompt Changed How I Use AI Forever
• When An AI Tool I Built Evaluated My Resume, I Learned What 100 Rejections Never Taught Me
  • Includes Python code that runs an AI resume evaluator

Use Mermaid to render charts. AIs are pretty good at it. GitHub and GitLab support it natively, which means your README.md or wiki can literally render the diagram itself, no screenshots required.

Copilot and CopilotChat install instructions are at the neovim page.

• Check out avante to emulate the behaviour of the Cursor AI IDE. It provides users with AI-driven code suggestions and the ability to apply these recommendations directly to their source files with minimal effort.
• and the supermaven plugin which integrates Supermaven. (An improved chat interface.)

Best setup: ssh + tmux (Prefix-b (to toggle a floating pane) or Prefix-c; p/n to nav windows) + vim

+----------------------+           +----------------------+
| vim    | vim         |           | LLM CLI              |
| (left) | (right)     |           |                      |
|        |             |           |                      |
|        |             |   tmux    |                      |
|        |             |  <----->  |                      |
|----------------------|           |                      |
| ls -1 | entr ...     |           |                      |
|                      |           |                      |
+----------------------+           +----------------------+

Here's a Review of 5 AI coding CLIs.

Recommended CLI flow:

1. Make it write a design doc you give feedback on first.
2. Let it run in yolo mode, but review the MR/PR just like you would with a junior dev.
   1. Both GitHub Copilot CLI and Claude Code support subagents, (make a code reviewer subagent). (Reddit Tip)
   2. Claude hooks for inter-agent communication.
3. Don't let it touch git commands aside from log and diff.
4. Use effective context engineering for AI agents from Anthropic.
   1. Manage your context window, clear or compact when necessary.
   2. Another post with the same idea: Never exceed 60% context. Split work into 4 phases: Research → Plan → Implement → Validate. Clear context between each.

Tools:

Note that projects like perfetto have a toplevel CLAUDE.md and GEMINI.md which each symlink to AGENTS.md.

• GitHub Copilot CLI: Should be easy to integrate for work, as we use Copilot already.
  • Customize with .copilot/instructions.
• Anthropic’s Claude-Code: “Unleash Claude’s raw power directly in your terminal.”
  • Customize with CLAUDE.md instructions.
• Google's Gemini CLI: Gemini CLI is an open-source AI agent that brings the power of Gemini directly into your terminal.
  • Customize with GEMINI.md.
• OpenAI’s Codex CLI a coding agent that you can run locally from your terminal and that can read, modify, and run code on your machine.
• OpenCode: An open source AI coding agent built for the terminal. (As seen here.)
  • Good conversation here.
  • Tip: Use existing Copilot subscription, and select Claude Sonnet 4.
• Charm Crush: At end of July 2025, Kujtim Hoxha left OpenCode and works at Charm.
• Aider: Aider lets you pair program with LLMs. Similar to OpenCode.
• Amp is an agentic coding tool available as a CLI and in VS Code, Cursor, and Windsurf.
  • Free Mode sessions are used for training.

A prompt to generate better prompts. Starts with:

You are Alisa, a master-level AI prompt optimization specialist. Your mission is to transform any user input into precision-crafted prompts that unlock the full potential of AI across all major platforms.

These 9 MCP Servers Improve AI Context (Reduce 99% Code Errors). This feels AI generated, but some of these MCP servers might be useful.

Keywords: LLM Large Language Model