Track and report OpenClaw model usage and costs. Generate daily/weekly cost reports with per-model token breakdowns, interactive vs cron classification, provider summary, and trend analysis.
Score and publish AI Feed candidates through a controlled runner. Agent only writes scored-results JSON.
Broadcast newly published AI Feed posts through a controlled wrapper with state and delivery guards.
AI Feed deterministic collection runner. Use to collect Miniflux/HN/GitHub candidates into data/candidates.json for feed-score.
Smart Discord thread archiving. Use when: (1) running periodic thread cleanup, (2) evaluating whether Discord threads should be archived. Agent lists active threads, reads messages, judges conversation status, archives resolved threads, and produces a…
技能实验室：OpenClaw 技能的全生命周期准入测试。包含三阶段： (1) 静态校验 — 结构、路径、跨平台兼容性（脚本 + AI 分析） (2) 智能分析 — 模型阅读技能全部内容，语义级审查 (3) 动态测试 — 生成子 Agent 加载技能实际运行，验证能力声明 Use when: 收到技能压缩包/目录需要评审，技能发布前准入测试，技能升级回归测试。
Publish Markdown articles to WeChat Official Account draft box using a local config file. Use when user wants to push blog posts to 微信公众号, convert Markdown to WeChat format, or create 公众号草稿.
Autonomous CC (Claude Code) iteration loop. Manages background coding agent tasks with pull-based issue tracking.