claude-tools

Provides a structured reference manual for Claude Code CLI tools, commands, configuration, and extensions.

10|3|Updated Mar 7, 2026
One-click install
npx skills add https://github.com/Cynthia1070711/PHYCOOL_Tools --skill claude-tools-cynthia1070711
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: claude-tools
Source: https://github.com/Cynthia1070711/PHYCOOL_Tools/tree/main/config-templates/claude/skills/claude-tools
Command: npx skills add https://github.com/Cynthia1070711/PHYCOOL_Tools --skill claude-tools-cynthia1070711

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? Claude Code has hundreds of features—30+ tools, 60+ commands, 50+ CLI flags, 80+ environment variables, hooks, MCP, plugins, and deployment options—scattered across dozens of official docs. This Skill consolidates them into a topic-indexed reference so you get accurate answers about Claude Code capabilities without searching documentation manually. ## Core Features & Use Cases - Topic-Based Lookup: Invoke with a topic (hooks, mcp, subagents, permissions, token-optimization, deployment, env-vars) to load only the relevant reference file. - Complete Coverage: References cover the extension system (Skills/Subagents/Agent Teams/MCP/Plugins/Hooks/Channels), configuration (settings priority, permission modes, sandboxing), automation (SDK mode, --bare, scheduling), cloud deployment (Bedrock/Vertex/Foundry), and 12 token optimization strategies. - PhyCool Cron Templates: Includes ready-to-use scheduled task templates (tech debt sweep, skill sync audit, token usage report) with [SILENT] notification semantics. - Use Case: When you ask "how do I configure permission rules for Bash commands" or "which environment variables control MCP timeouts", the Skill loads the matching reference and returns precise syntax and examples. ## Quick Start Ask the claude-tools skill to explain how hooks work and which events can block tool execution.

Frequently Asked Questions about claude-tools

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I configure permission rules in Claude Code?

Permission rules use patterns like Bash(npm run *) or Read(./secrets/**) in settings.json under permissions.allow, deny, or ask. Evaluation order is deny first, then ask, then allow, with the first match winning.

What hook events does Claude Code support?

Claude Code supports 25 hook events including SessionStart, UserPromptSubmit, PreToolUse, PostToolUse, Stop, and SessionEnd. Blocking-capable events like PreToolUse can exit with code 2 to prevent an action and feed stderr back to Claude.

How do I reduce token usage in Claude Code?

Use /clear between unrelated tasks, delegate research to subagents, move workflows into skills with disable-model-invocation, prefer CLI tools over MCP servers, and lower effort levels for simple tasks. The token-optimization reference lists 12 strategies across five categories.

Can Claude Code run on AWS Bedrock or Google Vertex?

Yes, set CLAUDE_CODE_USE_BEDROCK=1 or CLAUDE_CODE_USE_VERTEX=1 with the corresponding region and model environment variables. Always pin model versions with ANTHROPIC_DEFAULT_OPUS_MODEL and related variables when using cloud providers.

What is the difference between subagents and agent teams?

Subagents run in isolated contexts and report results back to the main conversation, while agent teams are fully independent Claude instances that coordinate via a shared task list and mailbox. Agent teams cost roughly 7x more tokens than a single agent.

Why are my MCP tools consuming too much context?

MCP tool schemas load into every request by default. Enable Tool Search with ENABLE_TOOL_SEARCH to defer schemas until needed, disable idle servers via /mcp, and cap responses with MAX_MCP_OUTPUT_TOKENS.