AgentSkillsCN

update-agent

创建、更新或管理 universal-ai-config 的代理模板。负责查找现有代理、判断是新建还是修改,并撰写模板。

SKILL.md
--- frontmatter
name: update-agent
description: Create, update, or manage universal-ai-config agent templates. Handles finding existing agents, deciding whether to create or modify, and writing the template.

Manage Agent Templates

Agents are specialized AI personas with scoped tools and permissions that run in isolated contexts.

Important: Agents are supported by Claude and Copilot only. Cursor does not support agents — consider using a skill with forkContext: true as an alternative for Cursor.

Finding Existing Agents

List files in .universal-ai-config/agents/ to discover existing agent templates. Read their frontmatter to understand each agent's purpose and capabilities.

Deciding What to Do

  • Create new: when you need a new specialized persona with distinct capabilities
  • Update existing: when an agent needs adjusted tools, permissions, or system prompt
  • Delete: when an agent is no longer needed

Creating a New Agent

  1. Create a .md file in .universal-ai-config/agents/ with a descriptive name (e.g. code-reviewer.md)
  2. Add YAML frontmatter with at minimum name and description
  3. Write the agent's system prompt as the body

Frontmatter Fields

FieldDescription
nameAgent identifier (lowercase with hyphens)
descriptionWhen to delegate to this agent (helps AI decide when to use it)
toolsTools this agent can use (omit to inherit all)
disallowedToolsTools to explicitly deny
permissionModePermission level: default, acceptEdits, dontAsk, bypassPermissions, plan
skillsSkills to preload into the agent's context
memoryPersistent memory scope: user, project, or local
modelAI model to use (e.g. sonnet, opus, haiku)
targetCopilot-specific: scope to vscode or github-copilot
mcpServersMCP server configurations
handoffsCopilot-specific: other agents to hand off to

Example

markdown
---
name: security-reviewer
description: Reviews code for security vulnerabilities and best practices. Use proactively after code changes.
tools: ["Read", "Grep", "Glob", "Bash"]
model: sonnet
---

You are a security-focused code reviewer. When invoked:

1. Identify the changed files
2. Check for common vulnerabilities (injection, XSS, auth issues)
3. Review dependency usage for known CVEs
4. Report findings by severity (critical, warning, info)

After Changes

Run uac generate to regenerate target-specific config files and verify the output.