Agent Authoring

Rules and patterns for writing agent definitions under claude/agents/. Use when creating or editing .md agent files.

Ven0m0 687aaeb 1.2 KB Updated

File contents

Agent Authoring

Agent definitions live in claude/agents/ as hyphenated-name.md files.

Required Frontmatter

---
name: kebab-case-name # matches filename without .md
description: one sentence # ≤512 chars; should make the agent self-select correctly
---

Body Structure

  1. Role — one sentence defining what this agent does and when to use it
  2. Capabilities — bullet list of specific things it can do
  3. Constraints — what it must NOT do (keeps scope tight)
  4. Instructions — step-by-step behavior, referencing tools by name
  5. Output format — how responses should be structured

Quality Rules

  • Description must be specific enough that an orchestrator picks it over another agent
  • Constraints section is mandatory — omitting it produces agents that overreach
  • Do not duplicate behavior already in a skill; reference the skill instead
  • Keep the whole file under 400 lines
  • No emojis anywhere in the file
  • Validate after writing: uv tool run "claudelint@0.3.3" --strict claude/agents/<name>.md

Ven0m0/claude-config/tree/main/.kilo/skill/agent-authoring commit 687aaebb7f

Frequently asked questions

npx skillmds@latest add ven0m0/agent-authoring