Back to directory

Claude Code System Prompts

A continuously updated extraction of every part of Claude Code's system prompt, including built-in tool descriptions and sub-agent prompts, with token counts and a changelog across hundreds of versions.

Claude Code System Prompts

This repository publishes an up-to-date extraction of Claude Code's system prompts: the descriptions behind its built-in tools, the separate prompts for its built-in sub-agents, and the prompts behind utility functions such as conversation compaction, project-memory-file generation, the status line, web fetching, shell-command handling, security review and agent creation, together with each prompt's token count for the current Claude Code release.

Because Claude Code assembles its behavior from more than 500 separate strings inside a large, minified JavaScript file rather than one fixed prompt, the project extracts them directly from the compiled source using a script, so the result matches what actually ships, and it also maintains a changelog tracking how these prompts have changed across roughly 300 prior versions of Claude Code.

A companion tool, mentioned in the same project, lets a user edit individual prompt pieces as separate Markdown files and then patch their own npm-based or native Claude Code installation with the changes, including diffing and conflict handling when both the user and a new Claude Code release have modified the same prompt.

More in Coding agents

See all 793
  • The City

    github.com

    A native macOS app that shows Claude Code runs as a city, with each project as a building, each team as a floor of robots, and every animation driven by a real event from the claude CLI's stream.

  • rig

    tryrig.ai

    A small, single-binary agent runtime that points at any OpenAI-compatible model, local or hosted, reads and edits a repository's code, runs its tests and remembers what it learned, with a loop implemented in a few hundred lines.

  • Jauvex

    jauvex.reindent.com

    An Electron desktop app that puts Claude Code, Codex and Grok coding-agent sessions side by side and lets you steer them by voice, with a small model handling fast routing decisions.