{"skill":{"slug":"hz-context-optimizer","displayName":"Huizai Context Optimizer","summary":"This skill should be used when the user asks to \"compress context\", \"summarize conversation history\", \"implement compaction\", \"reduce token usage\", or mentio...","tags":{"context":"1.0.0","latest":"1.0.0","openclaw":"1.0.0","optimization":"1.0.0","token":"1.0.0"},"stats":{"comments":0,"downloads":537,"installsAllTime":0,"installsCurrent":0,"stars":0,"versions":1},"createdAt":1771755074358,"updatedAt":1777525319712},"latestVersion":{"version":"1.0.0","createdAt":1771755074358,"changelog":"- Initial release of the context-compression skill for optimizing long-running agent sessions and large codebases.\n- Outlines when to activate context compression, including exceeding context window limits and cases where agents lose artifact tracking.\n- Describes three main compression strategies: Anchored Iterative Summarization, Opaque Compression, and Regenerative Full Summary, emphasizing structured summaries to prevent information loss.\n- Introduces the tokens-per-task metric and highlights the importance of minimizing total token usage, not just tokens per request.\n- Details the artifact trail problem and the necessity for explicit file tracking to support coding agent reliability.\n- Provides practical frameworks for triggering compression, evaluating quality, and integrating structured workflow phases for effective context management.","license":null},"metadata":null,"owner":{"handle":"lidekahdjdhdhsjjs-lang","userId":"publishers:lidekahdjdhdhsjjs-lang","displayName":"lidekahdjdhdhsjjs-lang","image":"https://avatars.githubusercontent.com/u/248733544?v=4"},"moderation":null}