品牌保护系统部署

部署 4 个智能体（agent）：scanner（扫描器）、classifier（分类器）、enforcer（执行器）和 dashboard（仪表盘）。

角色说明

部署流程

步骤 1： 向用户确认该智能体应承担的角色以及所使用的前缀（prefix）。

步骤 2： 安装所选角色对应的技能：

# scanner 使用：
clawhub install pilot-stream-data pilot-cron pilot-archive
# classifier 使用：
clawhub install pilot-event-filter pilot-task-router pilot-metrics
# enforcer 使用：
clawhub install pilot-webhook-bridge pilot-audit-log pilot-receipt
# dashboard 使用：
clawhub install pilot-metrics pilot-slack-bridge pilot-announce

步骤 3： 设置主机名：

pilotctl --json set-hostname -

步骤 4： 编写部署清单文件（setup manifest）：

mkdir -p ~/.pilot/setups
cat > ~/.pilot/setups/brand-protection.json << 'MANIFEST'

MANIFEST

步骤 5： 提示用户与直连通信对端（direct communication peers）发起握手（handshake）操作。

各角色对应的清单模板

scanner

{
  "setup": "brand-protection", "setup_name": "Brand Protection",
  "role": "scanner", "role_name": "Brand Scanner",
  "hostname": "-scanner",
  "description": "Crawls marketplaces, social media, and domains for unauthorized brand usage and counterfeits.",
  "skills": {"pilot-stream-data": "Stream crawled listings and social media posts for analysis.", "pilot-cron": "Schedule periodic scans across configured platforms.", "pilot-archive": "Archive raw scan results and evidence snapshots."},
  "peers": [{"role": "classifier", "hostname": "-classifier", "description": "Receives brand violations for classification"}],
  "data_flows": [{"direction": "send", "peer": "-classifier", "port": 1002, "topic": "brand-violation", "description": "Brand violations detected across platforms"}],
  "handshakes_needed": ["-classifier"]
}

classifier

{
  "setup": "brand-protection", "setup_name": "Brand Protection",
  "role": "classifier", "role_name": "Threat Classifier",
  "hostname": "-classifier",
  "description": "Classifies violations by type (counterfeit, impersonation, trademark), severity, and platform.",
  "skills": {"pilot-event-filter": "Filter noise and deduplicate violations across platforms.", "pilot-task-router": "Route violations to appropriate classification workflows by type.", "pilot-metrics": "Track violation counts, severity distribution, and platform breakdown."},
  "peers": [{"role": "scanner", "hostname": "-scanner", "description": "Sends brand violations"}, {"role": "enforcer", "hostname": "-enforcer", "description": "Receives classified threats"}],
  "data_flows": [{"direction": "receive", "peer": "-scanner", "port": 1002, "topic": "brand-violation", "description": "Brand violations detected across platforms"}, {"direction": "send", "peer": "-enforcer", "port": 1002, "topic": "classified-threat", "description": "Classified threats with type, severity, and priority"}],
  "handshakes_needed": ["-scanner", "-enforcer"]
}

enforcer

{
  "setup": "brand-protection", "setup_name": "Brand Protection",
  "role": "enforcer", "role_name": "Takedown Enforcer",
  "hostname": "-enforcer",
  "description": "Files DMCA notices, platform reports, and cease-and-desist requests. Tracks enforcement status.",
  "skills": {"pilot-webhook-bridge": "Submit takedown requests to platform APIs and legal services.", "pilot-audit-log": "Log all enforcement actions with evidence chain and timestamps.", "pilot-receipt": "Generate and store receipts for filed takedown notices."},
  "peers": [{"role": "classifier", "hostname": "-classifier", "description": "Sends classified threats"}, {"role": "dashboard", "hostname": "-dashboard", "description": "Receives enforcement actions"}],
  "data_flows": [{"direction": "receive", "peer": "-classifier", "port": 1002, "topic": "classified-threat", "description": "Classified threats to enforce"}, {"direction": "send", "peer": "-dashboard", "port": 1002, "topic": "enforcement-action", "description": "Enforcement actions with status and outcomes"}],
  "handshakes_needed": ["-classifier", "-dashboard"]
}

dashboard

{
  "setup": "brand-protection", "setup_name": "Brand Protection",
  "role": "dashboard", "role_name": "Brand Dashboard",
  "hostname": "-dashboard",
  "description": "Visualizes brand health metrics, violation trends, and enforcement success rates.",
  "skills": {"pilot-metrics": "Compute brand health scores, violation trends, and enforcement KPIs.", "pilot-slack-bridge": "Post brand health summaries and critical alerts to Slack.", "pilot-announce": "Broadcast periodic brand health reports to stakeholders."},
  "peers": [{"role": "enforcer", "hostname": "-enforcer", "description": "Sends enforcement actions"}],
  "data_flows": [{"direction": "receive", "peer": "-enforcer", "port": 1002, "topic": "enforcement-action", "description": "Enforcement actions with status and outcomes"}, {"direction": "send", "peer": "external", "port": 443, "topic": "brand-report", "description": "Brand health reports to stakeholders"}],
  "handshakes_needed": ["-enforcer"]
}

数据流向

• scanner -> classifier：来自平台扫描的 brand-violation 事件（端口 1002）
• classifier -> enforcer：含类型、严重程度与优先级的 classified-threat（端口 1002）
• enforcer -> dashboard：含状态与结果的 enforcement-action（端口 1002）
• dashboard -> external：通过 Slack 和 Webhook 发送的 brand-report（端口 443）

握手命令

# scanner <-> classifier：
pilotctl --json handshake -classifier "setup: brand-protection"
pilotctl --json handshake -scanner "setup: brand-protection"
# classifier <-> enforcer：
pilotctl --json handshake -enforcer "setup: brand-protection"
pilotctl --json handshake -classifier "setup: brand-protection"
# enforcer <-> dashboard：
pilotctl --json handshake -dashboard "setup: brand-protection"
pilotctl --json handshake -enforcer "setup: brand-protection"

工作流示例

# 在 scanner 上 —— 发布检测到的违规事件：
pilotctl --json publish -classifier brand-violation '{"violation_id":"BV-20481","platform":"amazon","type":"counterfeit","confidence":0.91}'
# 在 classifier 上 —— 发布已分类的威胁：
pilotctl --json publish -enforcer classified-threat '{"violation_id":"BV-20481","severity":"high","action":"dmca_takedown"}'
# 在 enforcer 上 —— 发布执法动作：
pilotctl --json publish -dashboard enforcement-action '{"violation_id":"BV-20481","action":"dmca_filed","status":"pending"}'
# 在 dashboard 上 —— 发布品牌报告：
pilotctl --json publish -dashboard brand-report '{"period":"2026-W15","violations":23,"takedowns_filed":18,"success_rate":0.78}'

依赖项

需安装 pilot-protocol 技能、pilotctl 可执行文件、clawhub 可执行文件，并确保后台守护进程（daemon）正在运行。