Playbook for Claude Code

Fix privilege escalation via harness allow-rules

Exploits target harness allow-rule approval. Review rules carefully to prevent broad access.

Guard against privilege escalation in coding agents

Adversaries can exploit permission updates to gain broad access. Review rules carefully.

Access Claude Code cloud sessions from your phone

Start and manage coding sessions remotely using your mobile device with Claude Code.

Connect GitHub to Claude Code cloud sessions

Integrate GitHub repositories with your Claude Code cloud sessions for seamless version control.

Run multiple parallel cloud sessions in Claude Code

Execute several coding environments simultaneously using Claude Code's parallel session feature.

Claude Code cloud sessions for remote coding

Run coding sessions in a cloud virtual machine, accessible from any device, even your phone.

Claude Code Projects setup and usage checklist

Steps to set up, manage threads, review pull requests, and create artifacts in Claude Code Projects.

Building a Hover-to-Reveal Email Mod for Claude Code

Step-by-step guide to creating a custom mod that reveals email addresses on hover.

Claude Code Mods for Customizing Claude Code Behavior

Custom functions that alter Claude Code's rendering, prompt inputs, and tool calls.

Build an app prototype with Claude design and code

Go from rough idea to clickable prototype using Claude's design and coding features.

Convert HTML animation to video using Claude Code

Stripe uses Claude Code for 55% of pull requests

Stripe orchestrates Claude Code on throwaway VMs to generate over half of its pull requests, improving efficiency.

Prevent LLM Agents from Deleting Execution Traces

External attackers and agents can delete traces; use independent logging to prevent this.

Default auto-compaction for agent context windows

Don't let models decide when to compact context; build agents that manage their own context window.

Guardrails for AI-generated production code

Essential checks and automated processes to ensure AI code quality and maintainability.

Apply higher standards to AI-generated code than human-written code.

AI code needs more rigorous checks than human code to ensure quality and maintainability.

Use grounded documentation to avoid invented APIs

Ensure AI agents use real-time, official documentation to prevent inventing deprecated APIs.

Google Cloud Developer agent plugin for AI coding

Simplifies AI-assisted cloud development by integrating docs, CLI tooling, and skill discovery.

Connect Claude Design MCP server to Claude Code

Connect the Claude Design MCP server to Claude Code to create and edit designs directly from your terminal.

Checklist for creating and testing custom Claude skills

Ensure your custom Claude skill is well-structured, functional, and correctly invoked.

Compliance requirements for government and regulated data

Checklist for handling CUI, ITAR, PHI, or PII data with Claude.

Claude product selection for government and regulated industries

Choose Claude for Government, Claude Enterprise, or Claude API via Bedrock/Vertex based on compliance needs.

Understanding Claude AI usage limits in Xcode

Claude usage limits are shared across platforms and reset every five hours.

Handling Claude AI rate limits in Xcode

If you hit a rate limit, wait for the reset period before trying again.

Focus custom skills on a single workflow

Design custom skills to perform one specific task rather than multiple unrelated functions.

Review and patch vulnerabilities in Claude Code

Use Claude Code to review proposed fixes for security vulnerabilities found by Claude Security.

Accessing Claude AI settings within Xcode

Configure Claude AI features by navigating to Xcode's preferences and Intelligence settings.

Enable artifacts in Claude for code execution and file creation

Turn on artifacts in Claude settings to allow code execution and file creation for enhanced tool building.

Enable code execution for custom skills

Ensure code execution is enabled to use custom skills that involve running code.

Manage automatic model switching in Claude Fable

Control if Claude automatically switches to a less capable model when Fable 5 or 5.1 is blocked.

Manage Claude Code authorization tokens

Remove authorization tokens for Claude Code to log out of that service.

Claude skill.md file structure and metadata

Template for the core skill.md file, including required metadata and optional content.

Claude AI coding assistance integrated into Xcode

Claude Sonnet 4 offers AI-powered coding help directly within Apple's Xcode IDE.

Claude artifacts for building shareable apps and tools

Artifacts let you turn ideas into shareable apps, tools, or content by describing what you need to Claude.

Claude Security scans code for vulnerabilities and suggests patches

Scans codebases for security flaws, identifies complex patterns, and proposes patches for review.

Use Claude as a pair programmer for coding tasks

Claude assists developers by acting as a pair programmer, proficient in major programming languages, especially Python.

Give Claude goals, not just tasks for coding

Frame coding requests as high-level objectives to leverage Claude's problem-solving capabilities.

Over-reliance on agent auto mode for security

Do not solely rely on agent auto mode for protection against prompt injection.

Claude Code as a universal harness for agentic environments

A component used in cloud-based agentic environments to provide a universal harness for interaction.

Success in agentic coding depends on problem understanding, not just coding skill

Understanding the problem you aim to solve is the key determinant of success with AI coding tools.

Claude Code for autonomous research analysis

Agentic coding platform that takes research ideas and datasets to write, run, and interpret analyses autonomously.

Prompt Claude Code to upgrade Anthropic SDK and pass tests

Prompt Claude Code to upgrade the Anthropic SDK and ensure tests pass after the upgrade.

Steps to regain control of AI coding workflows

Checklist for moving from cloud AI coding agents to local setups for better control.

Claude Code for AI-powered coding assistance

AI coding agent for generating and assisting with code development tasks.

Rebuild Python app with AI coding agents, from cloud to local

Learn to control AI coding workflows by rebuilding an app from cloud agents to local models.

Claude Code task setup for database cleanup

Defines a task for Claude Code to execute a database cleanup script with safety checks.

AI agent prevents data loss with voice confirmation

An AI agent used voice escalation to confirm database deletions, preventing data loss.

Voice AI data loss prevention demo highlights

Key steps in a demo showing AI preventing data loss via voice escalation.

Claude Code agent executes database deletion script

The agent executes a Python script for database cleanup after receiving confirmation.

AI coding agent resolves bug by calling developer for decisions

An AI coding agent used voice calls to resolve a checkout API bug, demonstrating human-AI collaboration.

AI agent calls developer for critical code fix decisions

AI agent autonomously resolves coding issues by calling developer for voice-based decisions on trade-offs.

Steps to get started with Claude Code locally

Checklist for installing Claude Code and setting up your local development environment.

Understand Claude Code permission modes

Grasp Claude Code's permission modes for controlled AI code generation and execution.

Set an autonomous goal for Claude Code

Use the /goal command to define and execute autonomous tasks with Claude Code.

Claude Code AI assistant for software development

AI assistant for software development to automate tasks and deploy applications.

Claude Agent Skills for task automation and expertise packaging

Package expertise into folders with instructions, scripts, and resources for Claude to load when needed.

Sonnet 4.6 vs. Opus 4.5 for coding tasks

Developers sometimes prefer Sonnet 4.6 over Opus 4.5 for coding due to better instruction following and less overengineering.

Sonnet 4.6 vs. Sonnet 4.5 for coding tasks

Sonnet 4.6 is preferred by users over Sonnet 4.5 for coding tasks due to better context reading and logic consolidation.

Use AI for comprehension, not just code generation

Leverage AI coding assistants to ask questions and seek explanations for deeper understanding.

Prioritize understanding over speed with AI coding tools

Focus on learning concepts when using AI coding assistants, not just task completion speed.

Claude Code auto mode defense layers

Understand the two defense layers in Claude Code's auto mode: input probe and output classifier.

Balance Claude Code safety and convenience with auto mode

Use auto mode to automate permission approvals for Claude Code, reducing fatigue while maintaining safety.

Tiered defenses in Claude Code auto mode

Understand the three tiers of defense in Claude Code's auto mode for evaluating actions.

Block arbitrary code execution tools in Claude Code auto mode

Disable broad code execution tools when entering auto mode to ensure the classifier sees potentially dangerous commands.

Agent oversight shifts from steps to strategy

Shift user oversight from individual actions to the overall plan for agent tasks.

CHANGELOG.md for agent progress tracking

Portable long-term memory for agents, tracking status, completed tasks, and failures.

CLAUDE.md for autonomous agent project planning

Project plan and design decisions for autonomous agents, editable by Claude.

Checklist for setting up autonomous scientific coding agents

Key items for configuring agents to work autonomously on scientific coding tasks.

Claude Code for agentic coding workflows

An AI coding assistant designed for complex, long-running agentic workflows.

Claude Code Checkpointing for Rewinding Code State

Automatically saves code state before changes, allowing rewinding to previous versions.

Reviewing AI-found vulnerabilities and suggested patches

Steps for human review of AI-identified security issues and proposed fixes.

AI finds subtle code vulnerabilities missed by traditional tools

AI can understand code context to find complex vulnerabilities beyond known patterns.

Relying on cloud AI with restrictive usage limits

Avoid cloud AI services with confusing token-based limits that hinder productivity for intensive coding tasks.

Goose vs. Claude Code for AI coding assistance

Goose offers free, local AI coding while Claude Code is a paid, cloud-based service with usage limits.

Develop a startup idea and implement it with AI

Generate a $1000/month startup idea and have AI implement it with no coding required.

Claude Code by Anthropic

An AI coding tool that generates ideas, writes code, and deploys applications autonomously.

Principles for writing high-quality AI agent tools

Key principles for creating effective tools that AI agents can use successfully.

Connect local MCP server to Claude Code

Command to register a local MCP server with Claude Code for tool testing.

Iteratively improve AI agent tools with Claude

Prototype, evaluate, and refine AI agent tools by collaborating with Claude.

Joint offering for CIOs to measure AI value and drive adoption

Structured path for CIOs to quantify AI productivity gains and ROI in engineering organizations.

Sandboxing reduces permission prompts and approval fatigue

Predefined sandbox boundaries allow Claude to work more freely, reducing constant permission requests.

Default Claude Code to read-only access

Claude Code operates on a permission-based model, defaulting to read-only access.

Enable Claude Code sandboxing for bash tool

Run /sandbox in Claude Code to enable sandboxing for the bash tool.

Claude Code command line tool for agentic coding

Command line tool for agentic coding, available as a limited research preview.

Claude AI integrated into Xcode for coding intelligence

Claude AI is now available within Xcode to enhance coding intelligence features for Apple platform development.

Claude Code for autonomous code development

An AI agent that autonomously writes, debugs, and edits code, used by software engineers.