Skills

agent-development

Installation
Summary

Design and build custom Claude Code agents with effective descriptions, tool access patterns, and self-documenting prompts.

  • Use strong description patterns with "MUST BE USED when" and "Use PROACTIVELY for" phrases to enable automatic task delegation; weak descriptions won't trigger auto-delegation
  • Restrict tool access to what agents actually need—omit Bash unless running scripts, and use allowlists in .claude/settings.json to reduce approval prompts
  • Encode all learnings directly into agent prompts using declarative instructions (describe what to accomplish, not how to use tools) so agents can reproduce behavior independently across sessions
  • Match specification level to task type: rigid steps for mechanical tasks, flexible guidelines for judgment-based work, and minimal constraints for creative tasks
  • Increase Node.js heap memory to 16GB via NODE_OPTIONS="--max-old-space-size=16384" to prevent memory crashes; limit parallel agents to 2-3 concurrent runs
SKILL.md

Agent Development for Claude Code

Build effective custom agents for Claude Code with proper delegation, tool access, and prompt design.

Agent Description Pattern

The description field determines whether Claude will automatically delegate tasks.

Strong Trigger Pattern

---
name: agent-name
description: |
  [Role] specialist. MUST BE USED when [specific triggers].
  Use PROACTIVELY for [task category].
  Keywords: [trigger words]
tools: Read, Write, Edit, Glob, Grep, Bash
model: sonnet
Related skills

More from jezweb/claude-skills

Installs
361
Repository
jezweb/claude-skills
GitHub Stars
776
First Seen
Jan 20, 2026
Security Audits
Gen Agent Trust HubWarn
SocketPass
SnykWarn