From agentic-ai-skills
Designs responsibility boundaries for business AI agents, defining Zone A (autonomous actions), Zone B (human confirmations), and Zone C (rejections/handoffs). Useful for quality interception, ticketing, and approval agent development.
npx claudepluginhub agenticaiplan/agenticaiskills --plugin agentic-ai-skillsThis skill uses the workspace's default tool permissions.
当用户需要为某个业务智能体明确以下内容时,使用本 Skill:
Defines roles for agents in multi-agent systems: purpose, capabilities, knowledge scope, authority, boundaries, success criteria. Covers principles, patterns, conflicts, and design artefacts.
Creates structured AI agent definitions with job titles, 15-30 domain vocabulary terms, deliverables, decision boundaries, SOPs, and anti-pattern watchlists. Use for defining specialized agents in any domain.
Designs and wires UiPath Human-in-the-Loop nodes for approval gates, escalations, write-back validation, and data enrichment in Flow, Maestro, or Coded Agent automations.
Share bugs, ideas, or general feedback.
当用户需要为某个业务智能体明确以下内容时,使用本 Skill:
典型业务场景包括:
输出必须使用结构化 Markdown,至少包含以下内容:
## Agent Boundary Spec
### 1. 智能体单一职责
- Job Story:
- 类型:建议型 / 执行型 / 混合型
### 2. Zone A:自主执行
- 动作:
- 触发条件:
- 调用系统(如适用):
### 3. Zone B:人工确认
| 动作 | 触发条件 | 确认方式 |
|------|----------|----------|
### 4. Zone C:拒绝执行并转人工
- 场景:
- 原因:
### 5. Handoff 规则
- 转人工条件:
- 转交内容:
- 恢复方式:
- 幂等保护:
- 审计要求:
### 6. 是否需要拆分多 Agent
- 结论:
- 理由:
### 7. 风险与待确认项
- 风险:
- 假设:
- 上线前检查项:
示例:发动机 100 阶啸叫质量拦截智能体
这个例子说明:边界规范不能只写“模型会不会”,而要写清楚采集条件、业务后果、接口动作、人工责任和审计链路。
如任务中涉及业务 SOP、合规规则、审批流、接口定义、角色权限或质量判定阈值,应优先结合对应资料一起判断边界,不要只根据抽象能力来划分 Agent 范围。