PropAI Live
Dependencies (auto-installed via ClawHub)
- - openclaw
- propai-sync
- social-flow
- message-parser
- lead-extractor
- india-location-normalizer
- summary-generator
- action-suggester
- sentiment-priority-scorer
- lead-storage
Core Workflow
For each session:
- 1. Validate environment: run
propai-live doctor. - Validate license: run
node scripts/license-status.mjs. - Parse intent: Lead (PropAI), Social (Social Flow), General (OpenClaw).
- Propose commands: show plan, confirm for writes.
- Execute minimal sequence, log outcomes.
- On failure, diagnose and retry safe actions only.
License Workflow
Run these from skills/propai-live:
CODEBLOCK0
CODEBLOCK1
Before any write operation:
CODEBLOCK2
Feature-specific gates:
CODEBLOCK3
For logout/device transfer:
CODEBLOCK4
Validate Environment
Before operations:
- -
propai-live --version (checks bundle) - INLINECODE4 (auth, connections)
- INLINECODE5 (license local check)
If issues are found, run propai-live setup for wizard.
Domain Routing
- - Lead/Messaging:
propai-live lead ... (PropAI + skills) - Social/Meta:
propai-live social ... (Social Flow) - AI/General:
propai-live ask ... (OpenClaw)
Execution Policy
- - Prefer CLI commands over narration.
- Keep commands profile-aware (agency, city).
- Do not print secrets in output.
- Run doctor first on new setups.
- Run
node scripts/license-guard.mjs --mode write before any write action.
Risk Policy
- - Read-only: safe (status, insights).
- Low-risk: drafts, test messages.
- High-risk: live spends, real user messages; require explicit confirmation.
- Unknown impact: treat as high-risk.
Output Contract
- - Summarize intent and chosen domain.
- List exact commands in code blocks.
- State assumptions (agency, contacts).
- Mark risk levels.
- Ask confirmation before writes.
References
- - PropAI docs: INLINECODE11
- Social Flow docs: INLINECODE12
- OpenClaw docs: INLINECODE13
- License API contract: INLINECODE14
- License API starter template: INLINECODE15
PropAI Live
依赖项（通过ClawHub自动安装）
- - openclaw
- propai-sync
- social-flow
- message-parser
- lead-extractor
- india-location-normalizer
- summary-generator
- action-suggester
- sentiment-priority-scorer
- lead-storage
核心工作流
针对每个会话：
- 1. 验证环境：运行 propai-live doctor。
- 验证许可证：运行 node scripts/license-status.mjs。
- 解析意图：线索（PropAI）、社交（Social Flow）、通用（OpenClaw）。
- 提出命令：显示计划，确认写入操作。
- 执行最小序列，记录结果。
- 失败时，诊断并仅重试安全操作。
许可证工作流
从 skills/propai-live 目录运行以下命令：
bash
node scripts/license-activate.mjs --key <许可证密钥> --api <许可证API基础URL>
bash
node scripts/license-status.mjs --mode read
在任何写入操作之前：
bash
node scripts/license-guard.mjs --mode write
功能特定门控：
bash
node scripts/license-guard.mjs --mode write --require-entitlement lead-storage
node scripts/license-guard.mjs --mode write --require-entitlement social
用于登出/设备转移：
bash
node scripts/license-deactivate.mjs
验证环境
操作之前：
- - propai-live --version（检查捆绑包）
- propai-live doctor（认证、连接）
- node scripts/license-status.mjs --mode read（许可证本地检查）
如果发现问题，运行 propai-live setup 启动向导。
域路由
- - 线索/消息：propai-live lead ...（PropAI + 技能）
- 社交/Meta：propai-live social ...（Social Flow）
- AI/通用：propai-live ask ...（OpenClaw）
执行策略
- - 优先使用CLI命令而非叙述。
- 保持命令配置文件感知（代理机构、城市）。
- 不在输出中打印机密信息。
- 在新设置上首先运行doctor。
- 在任何写入操作前运行 node scripts/license-guard.mjs --mode write。
风险策略
- - 只读：安全（状态、洞察）。
- 低风险：草稿、测试消息。
- 高风险：实时支出、真实用户消息；需要明确确认。
- 未知影响：视为高风险。
输出约定
- - 总结意图和所选域。
- 在代码块中列出确切命令。
- 说明假设条件（代理机构、联系人）。
- 标记风险等级。
- 在写入前请求确认。
参考资料
- - PropAI文档：propai-sync/README.md
- Social Flow文档：social-flow/README.md
- OpenClaw文档：https://openclaw.ai
- 许可证API合约：references/license-api-contract.md
- 许可证API启动模板：assets/license-api/
