Claude Code

Anthropic’s agentic coding tool. Runs in the terminal, reads your repo, and writes code with full context awareness.

Classifier-guarded permissions and a curated allow/deny hybrid for fast, safe sessions

Project-level startup instructions and conventions

Model selection, tool permissions, and runtime behavior

Reusable, composable instruction packs loaded on demand

Directory-scoped instructions with inheritance

Route Claude Code to third-party providers or local LLMs via env config