Claude Code

Claude Code is a terminal-based agentic coding assistant from Anthropic that operates within your local development environment. It reads and writes files, executes terminal commands, and connects to external tools via the Model Context Protocol (MCP).

Key Capabilities

• Codebase-aware: Reads your entire project context before making changes
• MCP-native: Connects to databases, APIs, trading platforms, and other tools
• Terminal-first: No IDE required — runs in your shell alongside your normal workflow
• Prompt-driven: Plan-before-code patterns produce more reliable results than freeform requests

Methodology

The vault includes two core methodology resources:

• claude-code-prompting-patterns — 10 prompting patterns from Anthropic's official workshop (Q&A-first, plan-then-build, feedback loops, CLAUDE.md, inline teaching)
• Claude Code settings — configuration reference including model routing via Kiro
• Codex Duo — field-tested workflow pairing Claude Code with Codex for adversarial code review, background rescue, and auto-review gating (see Codex)

See Also

• Codex — Claude Code's partner in the duo setup for code auditing and security review
• AI Agents & Automation MOC — full agent landscape, tool comparison, MCP patterns
• MCP — the protocol that connects Claude Code to external tools
• Safe AI Workflows - Mitigating Hallucinations and Overreach — safety protocols for agent work