Can I use this skill to build agents for specific engineering tasks?

Yes, it is specifically designed to help you write, build, and add new agent capabilities tailored to your specific codebase or workflow.

How do I distinguish between different agent types visually?

The skill provides a color convention system (e.g., Purple for Explore, Blue for Plan, Orange for Debug) set via YAML frontmatter.

Do I need to restart Claude Code after creating a new agent?

No, changes to agents take effect immediately through hot-reloading, allowing for rapid iteration.

What colors are available for custom agents?

While there are standard conventions, colors like Pink and Cyan are available for custom orchestration and coordination tasks.

What is Permission Denial Resilience?

It is a feature that allows subagents to automatically attempt alternative approaches if they hit a permission restriction, rather than stopping the workflow.

Agent Development Toolkit

Name: Agent Development Toolkit
Author: dwmkerr

bydwmkerr

•

Productivity & Workflow

Empowers developers to create, customize, and orchestrate specialized AI agents within Claude Code for automated engineering workflows.

The Agent Development skill streamlines the creation of specialized sub-agents in Claude Code, providing a standardized framework for defining roles, visual identities, and behaviors. It introduces a robust color-coding system to distinguish between planning, debugging, and execution agents, while implementing Permission Denial Resilience to ensure workflows continue autonomously even when encountering access restrictions. With support for instant hot-reloading, developers can iterate on agent capabilities without restarting their environment, making it an essential tool for scaling complex AI-driven development tasks.

Key Features

012 GitHub stars

02YAML-based agent configuration for simplified definition and metadata management

03Permission Denial Resilience for uninterrupted autonomous subagent workflows

04Instant hot-reloading for immediate application of agent modifications without restarts

05Specialized role assignments for analysis, security review, and orchestration

06Visual identity system using standardized color conventions for different agent roles

Use Cases

01Developing autonomous code exploration agents that navigate large repositories with resilient error handling

02Creating a multi-agent orchestration layer where a Plan agent coordinates with Create and Debug agents

03Building a specialized Security Review agent to perform automated audits during the development cycle

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add dwmkerr/claude-toolkit agent-development

For use in Claude.ai and ChatGPT

Key Features

012 GitHub stars

02YAML-based agent configuration for simplified definition and metadata management

03Permission Denial Resilience for uninterrupted autonomous subagent workflows

04Instant hot-reloading for immediate application of agent modifications without restarts

05Specialized role assignments for analysis, security review, and orchestration

06Visual identity system using standardized color conventions for different agent roles

Use Cases

01Developing autonomous code exploration agents that navigate large repositories with resilient error handling

02Creating a multi-agent orchestration layer where a Plan agent coordinates with Create and Debug agents

03Building a specialized Security Review agent to perform automated audits during the development cycle