Design and write Claude Code agents and prompts with delegation and orchestration details.
npx skills add https://github.com/inkeep/team-skills --skill write-agent
Please help me install this Agent Skill. Skill: write-agent Source: https://github.com/inkeep/team-skills/tree/main/plugins/shared/skills/write-agent Command: npx skills add https://github.com/inkeep/team-skills --skill write-agent
This Skill automates the complex process of designing and writing high-quality Claude Code agents and prompts, ensuring they are focused, safe, and reliable.
.claude/agents/ file.Use the write-agent skill to create a new agent that can review code changes for security vulnerabilities.
High-intent search queries and answers about installing and using this skill.
To design a Claude Code agent for workflow automation, you need to define its delegation triggers, tool permissions, and task-tool orchestration. This skill guides the creation of subagents and workflow orchestrators by generating ready-to-use agent files with optimized prompts.
A subagent is a single-purpose agent focused on a specific task, while a workflow orchestrator manages multi-phase processes. This skill helps you decide between them by addressing phase handoffs, iteration gates, and aggregation to ensure reliable agent development.
Structuring prompts for LLM agents requires defining clear output contracts and delegation triggers. This skill optimizes prompt engineering for clarity and reliability, ensuring your agents execute tasks safely within their defined scope and permissions.
You should use a subagent for focused delegation, a skill for reusable task execution, and always-on repo guidance for universal context. This skill facilitates the decision between subagents, skills, and repo guidance based on your specific automation needs.
Yes, you can generate a ready-to-use agent file for code review and security checks. This skill allows you to define the agent's purpose, scope, tools, and model choices, outputting a structured file for your `.claude/agents/` directory.
Limitations when creating multi-phase workflow orchestrators include managing complex phase handoffs and iteration gates without losing context. This skill addresses these constraints by enforcing output contracts and clear task-tool orchestration boundaries.
paradox123Collect repository structure, runnable commands, and architecture anchors at startup.
AJBcodingCoordinate multi-agent tasks and workflow orchestration in Claude Code environments.
yusufcmgImplement advanced prompt engineering patterns for LLM outputs using Python.
mituso89Execute development plans task by task through human-issued commands.