11. AI-assisted development
The practical art of building software with AI as a collaborator. Specifically: how to get the most out of Claude Code, the CLI agent I actually use day-to-day. Covers its features (memory, slash commands, hooks, plugins, subagents) and the prompt/workflow patterns that work.
Entries
| # | Entry | Status | One-line description |
|---|---|---|---|
| 1 | Claude Code deep dive | 🟩 COMPLETE 🟦 LIVING | The whole tour — what it is, what it can and can’t do, key concepts |
| 2 | The memory system | 🟩 COMPLETE 🟦 LIVING | CLAUDE.md, MEMORY.md, the memory/ folder — how persistent knowledge works |
| 3 | Slash commands | 🟩 COMPLETE 🟦 LIVING | /init, /help, /code-review, custom commands — extending Claude with shortcuts |
| 4 | Plugins | 🟩 COMPLETE 🟦 LIVING | Bundled commands + skills + connectors you can install |
| 5 | Hooks | 🟩 COMPLETE 🟦 LIVING | Shell commands that run on events (pre/post tool use, etc.) — automating behavior |
| 6 | settings.json | 🟩 COMPLETE 🟦 LIVING | The config file that controls permissions, env vars, hooks, model |
| 7 | Subagents | 🟩 COMPLETE 🟦 LIVING | Specialized agents (Explore, Plan, code-reviewer) spawned by the main agent |
| 8 | Workflow patterns | 🟩 COMPLETE | The day-to-day rhythms: plan → build → verify → commit |
| 9 | Prompt patterns for dev | 🟩 COMPLETE | Reusable prompt structures for code generation, review, debugging |
Suggested reading order
1 → 2 → 3 → 6 → 4 → 5 → 7 → 8 → 9. The first three are essential daily-use knowledge.
Vendor entries (coding agents, AI-first IDEs, agent frameworks)
Active build-out as of June 2026. Vendor entries follow the format from CONVENTIONS.md. Claude Code and its sub-features are documented in the entries above; this section catalogs alternatives and the broader landscape.
AI-first code editors:
| # | Entry | Status | One-line |
|---|---|---|---|
| 10 | 🇺🇸 Cursor | 🟩 🟦 | The most-used Claude Code alternative; VS Code fork rebuilt around AI |
| 11 | 🇺🇸 Windsurf (Codeium) | 🟩 🟦 | Direct Cursor competitor; previously Codeium Cascade |
| 12 | 🇺🇸 Zed AI | 🟩 🟦 | Rust-built fast editor with AI |
AI coding agents (terminal / CLI / agent-based):
| # | Entry | Status | One-line |
|---|---|---|---|
| 13 | 🇺🇸 OpenAI Codex (2026 agent) | 🟩 🟦 | OpenAI’s coding-agent answer to Claude Code |
| 14 | 🇺🇸 Google Antigravity | 🟩 🟦 | Google’s agentic coding product |
| 15 | 🇺🇸 Cline | 🟩 🟦 | Open-source VS Code agent |
| 16 | 🇺🇸 Roo Code | 🟩 🟦 | Open-source Cline-fork agent |
| 17 | 🇺🇸 Aider | 🟩 🟦 | Open-source CLI agent |
| 18 | 🇺🇸 Continue.dev | 🟩 🟦 | Open-source IDE agent framework |
| 19 | 🇺🇸 Devin (Cognition Labs) | 🟩 🟦 | Autonomous coding agent (full-task) |
IDE / editor plugins:
| # | Entry | Status | One-line |
|---|---|---|---|
| 20 | 🇺🇸 GitHub Copilot | 🟩 🟦 | The pioneer; multi-IDE; multi-model |
| 21 | 🇺🇸 GitHub Spark | 🟩 🟦 | GitHub’s “vibe coding” app builder |
| 22 | 🇨🇿 JetBrains AI Assistant | 🟩 🟦 | For IntelliJ / PyCharm / WebStorm |
| 23 | 🇮🇱 Tabnine | 🟩 🟦 | Privacy-focused completion |
| 24 | 🇺🇸 Amazon Q Developer | 🟩 🟦 | AWS-integrated (was CodeWhisperer) |
| 25 | 🇺🇸 Warp terminal AI | 🟩 🟦 | AI in the terminal itself |
| 26 | 🇫🇷 Mistral Codestral | 🟩 🟦 | French coding-specialist model |
| 27 | 🇺🇸 Microsoft Copilot Workspace | 🟩 🟦 | Microsoft’s project-level AI |
App-builder / “vibe coding” tools (sketch → working app):
| # | Entry | Status | One-line |
|---|---|---|---|
| 28 | 🇺🇸 v0 by Vercel | 🟩 🟦 | Generate React UIs from prompts; deploy to Vercel |
| 29 | 🇺🇸 Bolt.new (StackBlitz) | 🟩 🟦 | In-browser full-stack app generation |
| 30 | 🇸🇪 Lovable | 🟩 🟦 | Conversational app builder |
| 31 | 🇺🇸 Replit Agent | 🟩 🟦 | Replit’s app-builder agent |
| 32 | 🇺🇸 Sourcegraph Cody | 🟩 🟦 | Enterprise codebase-aware AI |
Anthropic-specific products (Phase D — beyond Claude Code core):
| # | Entry | Status | One-line |
|---|---|---|---|
| 33 | 🇺🇸 Claude Skills | 🟩 🟦 | Reusable instruction packs |
| 34 | 🇺🇸 Claude Cowork | 🟩 🟦 | Multi-instance collaborative mode |
| 35 | 🇺🇸 Claude Remote Desktop | 🟩 🟦 | Cloud-sandbox execution |
| 36 | 🇺🇸 Claude Remote Dispatch | 🟥 | Remote agent triggers |
| 37 | 🇺🇸 Anthropic Computer Use | 🟩 🟦 | Desktop agent (UI automation) |
| 38 | 🇺🇸 Claude in Chrome | 🟩 🟦 | Browser extension |
| 39 | 🇺🇸 Claude Agent SDK | 🟩 🟦 | Build agents with Anthropic SDK |
Agent frameworks (developer-facing):
| # | Entry | Status | One-line |
|---|---|---|---|
| 40 | LangGraph | 🟥 | The widely-used orchestration framework |
| 41 | 🇺🇸 LlamaIndex | 🟥 | RAG-focused framework |
| 42 | 🇺🇸 CrewAI | 🟥 | Multi-agent crews |
| 43 | 🇺🇸 Microsoft AutoGen | 🟥 | Microsoft’s multi-agent framework |
| 44 | 🇺🇸 Pydantic AI | 🟥 | Typed Python agents |
| 45 | 🇺🇸 OpenAI Agents SDK | 🟥 | OpenAI’s agent framework |
| 46 | 🇺🇸 Vercel AI SDK | 🟥 | Node / React AI framework |
Observability / evals:
| # | Entry | Status | One-line |
|---|---|---|---|
| 47 | 🇺🇸 LangSmith | 🟥 | LangChain’s observability |
| 48 | 🇩🇪 Langfuse | 🟥 | Open-source observability (Germany) |
| 49 | 🇺🇸 Helicone | 🟥 | LLM observability |
| 50 | 🇺🇸 LiteLLM | 🟥 | Router library |
| 51 | 🇺🇸 Portkey | 🟥 | AI Gateway |
| 52 | 🇺🇸 Braintrust | 🟥 | Evals platform |
| 53 | 🇺🇸 Promptfoo | 🟥 | Open-source evals |
| 54 | 🇬🇧 UK AISI Inspect AI | 🟥 | UK gov’s open-source evals tool |
See also
- 10. AI & LLMs 🟩 — the underlying tech
- AI Landscape — the map 🟩 🟦 — Western vs Chinese guidance, capability matrix
- 09. Tooling & dev environment 🟩 — where Claude Code fits among the other tools
- Decision frameworks — Cursor vs Claude Code vs Windsurf 🟥
- Claude Code official docs: https://docs.claude.com/en/docs/claude-code/overview