Modern Stack for AI-assisted coding

• IDE: Zed (over Cursor > PyCharm > VSCode)
  • Inline Auto-Complete: Zed Agent (> Github Copilot)
  • Inline Chat: Opus 4.5 (via Github Copilot)
  • Agentic Pannel: Claude Code tab (> OpenCode > Zed Agent)
    • Workflows
  • CLI: Claude Code / OpenCode (> Amp > Codex)
  • Git Commit Messages: Zed Agent (> Github Copilot)
  • Code Review: Claude Review (or Cursor BugBot)
  • Code Debugging: Cursor Debug with GPT5.2 High

• IDE/Terminal Panel Layout

  • Left Panel: File Explorer
  • Middle Panel: Current File
  • Right Panel: AI Agent open in Terminal or AI Panel

• Terminal:

  • tmux: for terminal panel management, and persistance of OpenCode/Claude sessions
  • oh-my-zsh: check out plugins – git, github cli, docker, docker-compose

• Python

  • Project Manager: uv
  • Language Server: Ruff, Ty
  • Python Environment: .venv
  • Environment/Config: .env -> config.py

• CLI Tools: for agent use during local development

  • rg (ripgrep): Fast grep searches for code navigation
  • gh CLI: GitHub Actions logs, issues, API
  • pixi run: Environment-aware Python execution
  • markdownlint: Validation
  • uvx commands: Package execution ** Include Local Norms in AGENTS.md **

• MCPs

  • (Frontend) Supabase, Firebase
  • External docs Context7

• Snippets

  • Whispr Flow: Examples cal link, work email

Thought Leaders on Agentic Coding

• https://www.anaconda.com/blog/challenging-claude-code-coding-competition

• Packages: https://github.com/StrongTheDev/logstash-for-zed https://github.com/mirageN1349/browser-tools-context-server

Agentic Development Cycle

Plan (design / review iterative loop) -> Execute -> Test -> Commit

Use cases

• Integrations (of external packages)