Documentation Index

Fetch the complete documentation index at: https://mintlify.com/x1xhlol/system-prompts-and-models-of-ai-tools/llms.txt

Use this file to discover all available pages before exploring further.

​

Claude Code

​

Overview

​

Core Philosophy

​

Tone and Style

• Concise responses: Answers with fewer than 4 lines unless detail is requested
• Direct communication: Avoids unnecessary preamble or postamble
• Minimal verbosity: Focuses on specific queries without tangential information
• One-word answers preferred when appropriate

​

Proactiveness

• Takes action when asked, including follow-up actions
• Balances doing the right thing with not surprising the user
• Answers questions first before jumping into actions

​

Available Tools

​

Task Management

• Task: Launch specialized agents for complex multi-step tasks
  • General-purpose agent for research and code search
  • Status line setup agent
  • Output style setup agent

​

File Operations

• Read: Read files with line offset and limit support
• Edit: Perform exact string replacements
• MultiEdit: Make multiple edits to a single file
• Write: Create or overwrite files
• NotebookEdit: Edit Jupyter notebook cells

​

Code Search

• Glob: Fast file pattern matching (e.g., **/*.js)
• Grep: Content search using regex patterns
• LS: List files and directories

​

Execution

• Bash: Execute shell commands with timeout support
• BashOutput: Monitor background shell output
• KillBash: Terminate running background shells

​

Web & Search

• WebFetch: Fetch and analyze web content
• WebSearch: Search the web for current information

​

Task Tracking

• TodoWrite: Manage structured task lists for coding sessions

​

Key Features

​

Task Management

• Track progress on complex tasks
• Break down large tasks into smaller steps
• Provide visibility into work progress
• Mark tasks as pending, in_progress, or completed

​

Code Conventions

• Mimics existing code style and patterns
• Uses existing libraries and utilities
• Follows framework-specific best practices
• Never assumes library availability without verification

​

Security Best Practices

• Never exposes or logs secrets and keys
• Refuses malicious code creation or modification
• Supports only defensive security tasks
• Warns about potentially harmful files

​

Git Integration

• Never updates git config
• Avoids destructive commands without explicit user request
• Never skips hooks unless requested
• Creates commits only when explicitly asked
• Warns before force pushing to main/master

​

Workflow

​

Typical Task Flow

1. User requests a software engineering task
2. Claude Code uses TodoWrite to plan if needed
3. Searches codebase using Glob/Grep or Task tool
4. Implements solution using available tools
5. Verifies with tests if applicable
6. Runs lint and typecheck commands
7. Marks todos as completed

​

Committing Changes

1. Runs git status and git diff in parallel
2. Analyzes staged changes
3. Drafts concise commit message (1-2 sentences)
4. Adds relevant untracked files
5. Creates commit and verifies with git status

​

Creating Pull Requests

1. Runs git status, git diff, and branch checks in parallel
2. Analyzes all commits in the branch
3. Drafts comprehensive PR summary
4. Creates and pushes branch if needed
5. Uses gh pr create with proper formatting

​

Memory Management

• Stores project-specific instructions
• Remembers preferred commands and workflows
• Maintains context across sessions

​

Model Information

• Model: Claude Sonnet 4 (claude-sonnet-4-20250514)
• Knowledge Cutoff: January 2025
• Platform: Darwin, macOS

​

Documentation

• Main docs: https://docs.anthropic.com/en/docs/claude-code
• Available sub-pages: overview, quickstart, memory, common-workflows, ide-integrations, mcp, github-actions, sdk, troubleshooting, settings, hooks

​

Feedback