agent-management

Use PROACTIVELY this agent when you need to design and create optimal Claude Code subagents, update existing agents with new capabilities, revise agent configurations, analyze project requirements to identify specialized roles, or craft precise agent configurations with appropriate tool permissions and model tiers. When the user specify "Create or Update subagent [name]", this skill must be triggered.

$ インストール

git clone https://github.com/Emz1998/avaris-ai /tmp/avaris-ai && cp -r /tmp/avaris-ai/.claude/skills/agents-management ~/.claude/skills/avaris-ai

// tip: Run this command in your terminal to install the skill


name: agent-management description: Use PROACTIVELY this agent when you need to design and create optimal Claude Code subagents, update existing agents with new capabilities, revise agent configurations, analyze project requirements to identify specialized roles, or craft precise agent configurations with appropriate tool permissions and model tiers. When the user specify "Create or Update subagent [name]", this skill must be triggered.

Goal: Create and maintain Claude Code subagents with appropriate tools, model tiers, and configurations.

IMPORTANT: Keep subagent content high-level and concise. Do not dive into implementation details.

Workflow

Phase 1: Assessment

  • Read .claude/skills/agents-management/references/subagent-doc.md
  • Read template at .claude/skills/agents-management/template.md
  • Analyze requirements and identify agent role
  • Check if agent exists (update vs create)
  • Determine model tier based on task complexity

Phase 2: Configuration

  • Define persona and core responsibilities
  • Select minimal required tool permissions
  • Structure workflow phases and constraints
  • Follow template structure exactly

Phase 3: Implementation

  • Write or update agent configuration file
  • Validate YAML frontmatter and sections
  • Save it to the appropriate folder in the .claude/agents directory
  • Report completion with agent details and location

Constraints

  • No unnecessary tool permissions
  • No duplicate or conflicting agent roles
  • Do not overengineer configurations
  • DO NOT deviate from the template structure

Acceptance Criteria

  • Agent file created/updated in .claude/agents/[team]/ folder
  • YAML frontmatter includes name, description, tools, model, color
  • Follows template structure with all required sections
  • No conflicts with existing agents in the ecosystem
  • Report delivered with location and usage guidance