{"skill":{"slug":"dragon-session-compactor","displayName":"Dragon Session Compactor","summary":"上下文压缩技能。当对话越来越长、token快爆的时候，自动压缩旧消息为摘要，保留最近上下文。 触发条件： - \"压缩会话\"、\"compact\"、\"上下文满了\" - 检测到消息超过阈值（默认100条或token估计超过80000） - 手动调用 $compact 无外部依赖，纯Node.js实现。","tags":{"compact":"1.0.0","context":"1.0.0","dragon":"1.0.0","latest":"1.0.0","session":"1.0.0","summary":"1.0.0"},"stats":{"comments":0,"downloads":80,"installsAllTime":0,"installsCurrent":0,"stars":0,"versions":1},"createdAt":1775482505797,"updatedAt":1775482911093},"latestVersion":{"version":"1.0.0","createdAt":1775482505797,"changelog":"Session Compactor v1.0.0\n\n- Initial release of a pure Node.js context compaction tool for long conversations.\n- Automatically summarizes old messages when conversation exceeds message or token limits (default: 100 messages or 80,000 tokens).\n- Supports manual and threshold-triggered compaction, with CLI commands for compacting, checking status, and viewing history.\n- Retains recent messages (default: last 4), extracts key info, and generates incremental conversation summaries.\n- No external dependencies; uses `.clawsession.json` for structured session storage.\n- Intended for OpenClaw and similar LLM chat systems to prevent context overflow.","license":"MIT-0"},"metadata":null,"owner":{"handle":"xhmqq616","userId":"s17evsy03dtqhcsv6fqy5yyqs184bmrr","displayName":"xhmqq616","image":"https://avatars.githubusercontent.com/u/230696037?v=4"},"moderation":{"isSuspicious":true,"isMalwareBlocked":false,"verdict":"suspicious","reasonCodes":["suspicious.llm_suspicious"],"summary":"Detected: suspicious.llm_suspicious","engineVersion":"v2.2.0","updatedAt":1775482911093}}