[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"extension-skill-thedotmack-mem-search-zh-CN":3,"guides-for-thedotmack-mem-search":466,"similar-k1780mdg10a6c0xrgsftrqjxg986mb1r-zh-CN":467},{"_creationTime":4,"_id":5,"children":6,"community":7,"display":9,"evaluation":15,"identity":258,"isFallback":239,"parentExtension":264,"providers":265,"relations":271,"repo":274,"tags":462,"workflow":463},1778698405106.5684,"k1780mdg10a6c0xrgsftrqjxg986mb1r",[],{"reviewCount":8},0,{"description":10,"installMethods":11,"name":13,"sourceUrl":14},"搜索 Claude-Mem 的持久化跨会话记忆数据库。当用户问“我们是否已经解决了这个问题？”、“上次我们如何做 X？”或者需要处理先前会话中的工作时使用。",{"claudeCode":12},"thedotmack/claude-mem","Claude-Mem","https://github.com/thedotmack/claude-mem",{"_creationTime":16,"_id":17,"extensionId":5,"locale":18,"result":19,"trustSignals":237,"workflow":256},1778698405106.5686,"kn7dsqmg897bszt5p0gpxcr8nn86m9qy","zh-CN",{"checks":20,"evaluatedAt":191,"extensionSummary":192,"features":193,"nonGoals":199,"practices":204,"prerequisites":208,"promptVersionExtension":213,"promptVersionScoring":214,"purpose":215,"rationale":216,"score":217,"summary":218,"tags":219,"tier":225,"useCases":226,"workflow":231},[21,26,29,32,36,39,43,47,50,53,57,61,64,68,71,74,77,80,83,86,90,94,98,102,106,109,112,115,119,122,125,128,131,134,137,141,145,149,153,157,160,163,166,169,173,176,179,182,185,188],{"category":22,"check":23,"severity":24,"summary":25},"实际效用","问题相关性","pass","描述清楚地说明了记住过去工作和会话的问题，并提供了具体的示例用户问题。",{"category":22,"check":27,"severity":24,"summary":28},"独特卖点","该扩展提供了持久化记忆数据库以及用于搜索和检索过去工作的特定工具，超越了基本的 LLM 上下文。",{"category":22,"check":30,"severity":24,"summary":31},"生产就绪","该扩展为搜索、过滤和获取过去观察结果提供了一个完整的工作流，涵盖了检索历史上下文的整个生命周期。",{"category":33,"check":34,"severity":24,"summary":35},"范围","单一职责原则","该扩展仅专注于持久化记忆搜索和检索，不引入无关的功能。",{"category":33,"check":37,"severity":24,"summary":38},"描述质量","显示的描述准确地反映了 SKILL.md 和 README 中描述的扩展功能。",{"category":40,"check":41,"severity":24,"summary":42},"调用","作用域工具","该扩展公开了狭窄的动词-名词工具（`search`、`timeline`、`get_observations`）用于特定的记忆操作。",{"category":44,"check":45,"severity":24,"summary":46},"文档","配置和参数参考","MCP 工具的所有参数都在 SKILL.md 中得到了清楚的文档记录，包括适用的默认值。",{"category":33,"check":48,"severity":24,"summary":49},"工具命名","工具名称（`search`、`timeline`、`get_observations`）具有描述性，并遵循动词-名词约定。",{"category":33,"check":51,"severity":24,"summary":52},"最小 I/O 接口","工具参数是具体的且有文档记录，响应专注于请求的数据，没有过多的诊断信息。",{"category":54,"check":55,"severity":24,"summary":56},"许可证","许可证可用性","该扩展根据 Apache-2.0 许可证授权，如 LICENSE 文件和 README 中所示。",{"category":58,"check":59,"severity":24,"summary":60},"维护","提交的近况","存储库显示了最近的提交，表明正在积极维护。",{"category":58,"check":62,"severity":24,"summary":63},"依赖管理","该项目使用 `package.json` 和 `npm`，并且安装说明建议使用 `npx claude-mem install`，这表明标准的依赖管理实践。自动安装处理依赖项。",{"category":65,"check":66,"severity":24,"summary":67},"安全","秘密管理","该扩展在其核心功能中似乎不处理或暴露秘密；它操作用户提供的数据和元数据。",{"category":65,"check":69,"severity":24,"summary":70},"注入","该扩展的工具操作结构化查询和 ID，并且 SKILL.md 强调将搜索结果视为数据，从而降低了注入风险。",{"category":65,"check":72,"severity":24,"summary":73},"传递性供应链炸弹","安装过程使用标准的 `npx`，并且代码在运行时似乎没有获取或执行任意远程代码。",{"category":65,"check":75,"severity":24,"summary":76},"沙箱隔离","该扩展在 Claude Code 环境内运行，并与其数据库和工具进行交互；没有证据表明其写入超出指定范围或写入敏感系统路径。",{"category":65,"check":78,"severity":24,"summary":79},"沙箱逃逸原语","在提供的代码片段中未发现分离的进程生成或围绕被拒绝的工具调用的重试循环。",{"category":65,"check":81,"severity":24,"summary":82},"数据渗漏","该扩展的目的是在用户自己的环境中存储和检索用户过去的 work，没有迹象表明将此数据发送给第三方。",{"category":65,"check":84,"severity":24,"summary":85},"隐藏文本技巧","捆绑的文件和文档似乎没有隐藏文本技巧或恶意指令。",{"category":87,"check":88,"severity":24,"summary":89},"钩子","不透明代码执行","代码似乎是标准的 TypeScript/JavaScript，没有证据表明存在混淆、base64 载荷或运行时脚本获取。",{"category":91,"check":92,"severity":24,"summary":93},"可移植性","结构假设","该扩展假定存在标准的 Claude Code 环境和工具可用性，但不对其自身范围之外的用户项目文件结构做任何假设。",{"category":95,"check":96,"severity":24,"summary":97},"信任","问题关注度","与开放问题相比，存储库中的问题关闭率很高，这表明维护者响应迅速。",{"category":99,"check":100,"severity":24,"summary":101},"版本控制","发布管理","该扩展在 README 和 `package.json` 中有一个清晰的版本号 (6.5.0)，并且安装方法允许版本固定。",{"category":103,"check":104,"severity":24,"summary":105},"代码执行","验证","工具参数有文档记录且具体，这意味着在 MCP 框架内进行了验证。工作流强调在获取之前进行过滤，这表明了一定程度的数据验证。",{"category":65,"check":107,"severity":24,"summary":108},"无保护的破坏性操作","该扩展主要对用户数据是只读的；其操作涉及搜索和检索，而不是破坏性操作。",{"category":103,"check":110,"severity":24,"summary":111},"错误处理","SKILL.md 描述了一个结构化的工作流，并强调在获取详细信息之前进行过滤，这暗示了健壮的错误处理以防止过多的 API 调用。",{"category":103,"check":113,"severity":24,"summary":114},"日志记录","文档提到了日志记录设置和用于审查操作的审计日志功能，表明了适当的日志记录实践。",{"category":116,"check":117,"severity":24,"summary":118},"合规性","GDPR","该扩展操作用户过去的 work 和会话数据，而不是来自外部来源的个人数据。可以通过私有标签排除敏感内容。",{"category":116,"check":120,"severity":24,"summary":121},"目标市场","该扩展专为与 Claude Code 和其工具一起使用而设计，使其在全球范围内适用，没有特定的区域限制。",{"category":91,"check":123,"severity":24,"summary":124},"运行时稳定性","该扩展针对 Claude Code 环境和标准的 Node.js/Bun 运行时，并提供了 Windows 的清晰设置说明，表明了跨平台兼容性。",{"category":44,"check":126,"severity":24,"summary":127},"README","README 内容全面、结构良好，并清楚地说明了扩展的目的、功能和用法。",{"category":33,"check":129,"severity":24,"summary":130},"工具接口大小","该扩展公开了三个核心 MCP 工具，这是一个可管理且集中的接口区域。",{"category":40,"check":132,"severity":24,"summary":133},"重叠的近义词工具","这些工具（`search`、`timeline`、`get_observations`）具有不同的目的，并且似乎没有作为近义词重叠。",{"category":44,"check":135,"severity":24,"summary":136},"幻影功能","所有宣传的功能，包括持久化记忆、搜索工具和工作流，都已实现并记录在案。",{"category":138,"check":139,"severity":24,"summary":140},"安装","安装说明","为各种环境（npx、Gemini CLI、OpenCode、插件市场）提供了清晰、可复制粘贴的安装说明，并附有验证步骤。",{"category":142,"check":143,"severity":24,"summary":144},"错误","可操作的错误消息","虽然没有详细说明具体的错误消息格式，但结构化的工作流和文档暗示了清晰的错误报告和恢复指南。",{"category":146,"check":147,"severity":24,"summary":148},"执行","固定的依赖项","使用 `package.json` 和 `npx` 进行安装意味着依赖项已由 npm/Bun 固定管理。",{"category":33,"check":150,"severity":151,"summary":152},"干运行预览","not_applicable","该扩展主要用于检索信息，不执行需要干运行模式的破坏性操作。",{"category":154,"check":155,"severity":24,"summary":156},"协议","幂等重试和超时","工作流强调在获取之前进行过滤和批量请求，这表明了一个支持幂等性和高效重试的设计。具体的超时时间未详细说明，但通常由底层 MCP 框架管理。",{"category":116,"check":158,"severity":24,"summary":159},"遥测选择加入","文档提到了日志记录和数据处理的配置，暗示了对遥测的控制，并且没有发现选择退出的遥测。",{"category":40,"check":161,"severity":24,"summary":162},"精确目的","描述清楚地说明了目的（搜索记忆数据库）和使用场景（询问过去的工作、之前的会话），并有明确的触发器。",{"category":40,"check":164,"severity":24,"summary":165},"简洁的前言","SKILL.md 中的前言简洁明了，有效地总结了扩展的核心功能和目的。",{"category":44,"check":167,"severity":24,"summary":168},"简洁正文","SKILL.md 结构良好，并将详细信息委托给外部文档链接，保持主文件简洁。",{"category":170,"check":171,"severity":24,"summary":172},"上下文","渐进式披露","SKILL.md 概述了主要工作流，并链接到完整文档以进行深入研究，展示了渐进式披露。",{"category":170,"check":174,"severity":151,"summary":175},"分支探索","此技能侧重于查询现有数据，不涉及需要分支上下文的深入探索。",{"category":22,"check":177,"severity":24,"summary":178},"使用示例","SKILL.md 和 README 中提供了多个清晰、端到端的示例，演示了调用和预期结果。",{"category":22,"check":180,"severity":24,"summary":181},"边缘情况","SKILL.md 概述了一个带有参数和潜在过滤器的清晰工作流，暗示了对各种搜索场景的处理，并且文档涵盖了故障排除。",{"category":103,"check":183,"severity":151,"summary":184},"工具回退","该扩展依赖于 Claude Code 的内部 MCP 工具，不依赖需要回退的外部 MCP 服务器。",{"category":65,"check":186,"severity":24,"summary":187},"意外状态下停止","结构化的工作流和强调在获取之前进行过滤表明，意外状态会以清晰的报告优雅地停止进程。",{"category":91,"check":189,"severity":24,"summary":190},"跨技能耦合","该技能独立操作其记忆数据库，并且似乎不依赖于其他特定技能的并发加载。",1778698211503,"该扩展为 Claude Code 提供了一个持久化的跨会话记忆数据库，使用户能够搜索和检索过去的工作、观察和提示。它使用三层 MCP 工具工作流（搜索、时间线、获取观察结果）来高效查询历史数据，并包含项目过滤和时间顺序上下文检索等功能。",[194,195,196,197,198],"持久化跨会话记忆存储","高级搜索和过滤功能","观察结果周围的时间线检索","通过渐进式披露工作流实现高效的 token 使用","与 Claude Code 和 MCP 工具集成",[200,201,202,203],"充当通用任务管理器","存储实时对话上下文","替换主要的 Claude Code 对话历史记录","提供项目管理或错误跟踪功能",[205,206,207],"记忆管理","搜索索引","上下文检索",[209,210,211,212],"Node.js 18.0.0 或更高版本","支持插件的 Claude Code","Bun（如果缺失则自动安装）","uv（如果缺失则自动安装）","3.0.0","4.4.0","让 Claude Code 用户能够轻松回忆和利用不同会话中的过去工作和知识，从而提高连续性并减少重复工作。","该扩展的文档非常齐全且实用，目标明确，工具定义清晰，并附有强大的示例。它遵循安全、可维护和用户体验的最佳实践，因此获得了非常高的信任评分。",99,"一个高质量的持久化记忆搜索扩展，提供清晰的实用性和强大的实现。",[220,221,222,223,224],"memory","search","context","persistence","claude-code","verified",[227,228,229,230],"回忆先前遇到问题的解决方案","审查过去的项目决策和发现","查找先前工作会话中的特定信息","在长期项目中保持上下文",[232,233,234,235,236],"使用 `search` 工具搜索记忆索引以获取相关观察结果。","按查询、类型、日期、项目等过滤结果。","使用 `timeline` 工具获取有趣结果周围的时间顺序上下文。","根据搜索和时间线上下文选择相关的观察结果 ID。","使用 `get_observations` 工具获取选定观察结果的完整详细信息。",{"codeQuality":238,"collectedAt":240,"documentation":241,"maintenance":244,"popularity":250,"security":252,"testCoverage":255},{"hasLockfile":239},false,1778698195953,{"descriptionLength":242,"readmeSize":243},178,17010,{"closedIssues90d":245,"forks":246,"hasChangelog":239,"openIssues90d":247,"pushedAt":248,"stars":249},781,6488,53,1778643518000,75457,{"npmDownloads":251},124555,{"hasNpmPackage":253,"license":254,"smitheryVerified":239},true,"Apache-2.0",{"hasCi":253,"hasTests":253},{"updatedAt":257},1778698405106,{"basePath":259,"githubOwner":260,"githubRepo":261,"locale":18,"slug":262,"type":263},"plugin/skills/mem-search","thedotmack","claude-mem","mem-search","skill",null,{"evaluate":266,"extract":269},{"promptVersionExtension":213,"promptVersionScoring":214,"score":217,"tags":267,"targetMarket":268,"tier":225},[220,221,222,223,224],"global",{"commitSha":270,"license":254},"HEAD",{"repoId":272,"translatedFrom":273},"kd70jnxgm695az2wtf37zbqdj986mp7k","k1790kka1gerjkd8t2y1bfct6586ne6s",{"_creationTime":275,"_id":272,"identity":276,"providers":277,"workflow":457},1778698049260.0457,{"githubOwner":260,"githubRepo":261,"sourceUrl":14},{"classify":278,"discover":429,"extract":432,"github":433,"npm":456},{"commitSha":270,"extensions":279},[280,292,305,319,328,336,347,355,363,371,377,385,393,401,412,420],{"basePath":281,"displayName":282,"installMethods":283,"rationale":284,"selectedPaths":285,"source":289,"sourceLanguage":290,"type":291},".agents/plugins","claude-mem-local",{"claudeCode":12},"marketplace.json at .agents/plugins/marketplace.json",[286],{"path":287,"priority":288},"marketplace.json","mandatory","rule","en","marketplace",{"basePath":293,"description":294,"displayName":260,"installMethods":295,"rationale":296,"selectedPaths":297,"source":289,"sourceLanguage":290,"type":291},"","Plugins by Alex Newman (thedotmack)",{"claudeCode":12},"marketplace.json at .claude-plugin/marketplace.json",[298,300,302],{"path":299,"priority":288},".claude-plugin/marketplace.json",{"path":301,"priority":288},"README.md",{"path":303,"priority":304},"LICENSE","high",{"basePath":293,"description":306,"displayName":261,"installMethods":307,"license":254,"rationale":308,"selectedPaths":309,"source":289,"sourceLanguage":290,"type":318},"Memory compression system for Claude Code - persist context across sessions",{"claudeCode":261},"plugin manifest at .claude-plugin/plugin.json (coalesced with duplicate plugin at plugin)",[310,312,313,314,316],{"path":311,"priority":288},".claude-plugin/plugin.json",{"path":301,"priority":288},{"path":303,"priority":304},{"path":315,"priority":288},".mcp.json",{"path":317,"priority":304},".codex-plugin/plugin.json","plugin",{"basePath":320,"description":321,"displayName":322,"installMethods":323,"rationale":324,"selectedPaths":325,"source":289,"sourceLanguage":290,"type":263},"plugin/skills/babysit","Watch a pull request or review cycle until it is ready to merge. Use when asked to babysit, monitor, or keep checking PR comments, reviews, and CI until all actionable issues are resolved.","babysit",{"claudeCode":12},"SKILL.md frontmatter at plugin/skills/babysit/SKILL.md",[326],{"path":327,"priority":288},"SKILL.md",{"basePath":329,"description":330,"displayName":331,"installMethods":332,"rationale":333,"selectedPaths":334,"source":289,"sourceLanguage":290,"type":263},"plugin/skills/do","Execute a phased implementation plan using subagents. Use when asked to execute, run, or carry out a plan — especially one created by make-plan.","do",{"claudeCode":12},"SKILL.md frontmatter at plugin/skills/do/SKILL.md",[335],{"path":327,"priority":288},{"basePath":337,"description":338,"displayName":339,"installMethods":340,"rationale":341,"selectedPaths":342,"source":289,"sourceLanguage":290,"type":263},"plugin/skills/how-it-works","Explain how claude-mem captures observations, when memory injection kicks in, and where data lives. Use when the user asks \"how does claude-mem work?\" or \"what is this thing doing?\".","how-it-works",{"claudeCode":12},"SKILL.md frontmatter at plugin/skills/how-it-works/SKILL.md",[343,344],{"path":327,"priority":288},{"path":345,"priority":346},"onboarding-explainer.md","medium",{"basePath":348,"description":349,"displayName":350,"installMethods":351,"rationale":352,"selectedPaths":353,"source":289,"sourceLanguage":290,"type":263},"plugin/skills/knowledge-agent","Build and query AI-powered knowledge bases from claude-mem observations. Use when users want to create focused \"brains\" from their observation history, ask questions about past work patterns, or compile expertise on specific topics.","knowledge-agent",{"claudeCode":12},"SKILL.md frontmatter at plugin/skills/knowledge-agent/SKILL.md",[354],{"path":327,"priority":288},{"basePath":356,"description":357,"displayName":358,"installMethods":359,"rationale":360,"selectedPaths":361,"source":289,"sourceLanguage":290,"type":263},"plugin/skills/learn-codebase","Prime a codebase by reading every source file in full. Use when starting work on a new or unfamiliar project, or when the user asks to \"learn the codebase\", \"read the codebase\", \"prime\", or \"get up to speed\".","learn-codebase",{"claudeCode":12},"SKILL.md frontmatter at plugin/skills/learn-codebase/SKILL.md",[362],{"path":327,"priority":288},{"basePath":364,"description":365,"displayName":366,"installMethods":367,"rationale":368,"selectedPaths":369,"source":289,"sourceLanguage":290,"type":263},"plugin/skills/make-plan","Create a detailed, phased implementation plan with documentation discovery. Use when asked to plan a feature, task, or multi-step implementation — especially before executing with do.","make-plan",{"claudeCode":12},"SKILL.md frontmatter at plugin/skills/make-plan/SKILL.md",[370],{"path":327,"priority":288},{"basePath":259,"description":372,"displayName":262,"installMethods":373,"rationale":374,"selectedPaths":375,"source":289,"sourceLanguage":290,"type":263},"Search claude-mem's persistent cross-session memory database. Use when user asks \"did we already solve this?\", \"how did we do X last time?\", or needs work from previous sessions.",{"claudeCode":12},"SKILL.md frontmatter at plugin/skills/mem-search/SKILL.md",[376],{"path":327,"priority":288},{"basePath":378,"description":379,"displayName":380,"installMethods":381,"rationale":382,"selectedPaths":383,"source":289,"sourceLanguage":290,"type":263},"plugin/skills/pathfinder","Map a codebase into feature-grouped flowcharts, identify duplicated concerns across features, and propose a unified architecture. Use when asked to \"find the ideal path,\" unify duplicated systems, or audit architecture before a refactor. Emits a proposed unified flowchart plus per-system /make-plan prompts.","pathfinder",{"claudeCode":12},"SKILL.md frontmatter at plugin/skills/pathfinder/SKILL.md",[384],{"path":327,"priority":288},{"basePath":386,"description":387,"displayName":388,"installMethods":389,"rationale":390,"selectedPaths":391,"source":289,"sourceLanguage":290,"type":263},"plugin/skills/smart-explore","Token-optimized structural code search using tree-sitter AST parsing. Use instead of reading full files when you need to understand code structure, find functions, or explore a codebase efficiently.","smart-explore",{"claudeCode":12},"SKILL.md frontmatter at plugin/skills/smart-explore/SKILL.md",[392],{"path":327,"priority":288},{"basePath":394,"description":395,"displayName":396,"installMethods":397,"rationale":398,"selectedPaths":399,"source":289,"sourceLanguage":290,"type":263},"plugin/skills/timeline-report","Generate a \"Journey Into [Project]\" narrative report analyzing a project's entire development history from claude-mem's timeline. Use when asked for a timeline report, project history analysis, development journey, or full project report.","timeline-report",{"claudeCode":12},"SKILL.md frontmatter at plugin/skills/timeline-report/SKILL.md",[400],{"path":327,"priority":288},{"basePath":402,"description":403,"displayName":404,"installMethods":405,"rationale":406,"selectedPaths":407,"source":289,"sourceLanguage":290,"type":263},"plugin/skills/version-bump","Automated semantic versioning and release workflow for Claude Code plugins. Handles version increments across package.json, marketplace.json, plugin.json manifests, npm publishing (so `npx claude-mem@X.Y.Z` resolves), build verification, git tagging, GitHub releases, and changelog generation.","claude-code-plugin-release",{"claudeCode":12},"SKILL.md frontmatter at plugin/skills/version-bump/SKILL.md",[408,409],{"path":327,"priority":288},{"path":410,"priority":411},"scripts/generate_changelog.js","low",{"basePath":413,"description":414,"displayName":415,"installMethods":416,"rationale":417,"selectedPaths":418,"source":289,"sourceLanguage":290,"type":263},"plugin/skills/wowerpoint","Turn one document into a kawaii NotebookLM slide-deck PDF. Use for \"wowerpoint this\", \"make a deck about \u003Cfile>\", \"turn this report into slides\", or any request to render a single document as shareable narrative slides.","wowerpoint",{"claudeCode":12},"SKILL.md frontmatter at plugin/skills/wowerpoint/SKILL.md",[419],{"path":327,"priority":288},{"basePath":293,"description":306,"displayName":261,"installMethods":421,"license":254,"rationale":422,"selectedPaths":423,"source":289,"sourceLanguage":290,"type":428},{"npm":261},"package.json @modelcontextprotocol/sdk dep + bin at package.json",[424,426,427],{"path":425,"priority":288},"package.json",{"path":301,"priority":288},{"path":303,"priority":304},"mcp",{"sources":430},[431],"manual",{"npmPackage":261},{"closedIssues90d":245,"description":434,"forks":246,"homepage":435,"license":254,"openIssues90d":247,"pushedAt":248,"readmeSize":243,"stars":249,"topics":436},"Persistent Context Across Sessions for Every Agent –  Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More","https://claude-mem.ai",[437,438,439,440,441,442,443,444,224,445,446,447,448,449,450,451,452,453,454,455],"ai","ai-agents","ai-memory","anthropic","artificial-intelligence","claude","claude-agent-sdk","claude-agents","claude-code-plugin","claude-skills","long-term-memory","mem0","openmemory","supermemory","chromadb","sqlite","embeddings","memory-engine","rag",{"downloads":251},{"classifiedAt":458,"discoverAt":459,"extractAt":460,"githubAt":460,"npmAt":461,"updatedAt":458},1778698055851,1778698049260,1778698052709,1778698054459,[224,222,220,223,221],{"evaluatedAt":464,"extractAt":465,"updatedAt":257},1778698211608,1778698056313,[],[468,499,529,560,591,619],{"_creationTime":469,"_id":470,"community":471,"display":472,"identity":478,"providers":481,"relations":490,"tags":494,"workflow":495},1778694473202.7083,"k17fabzte0hynzp1fmsvfh1en986nzex",{"reviewCount":8},{"description":473,"installMethods":474,"name":476,"sourceUrl":477},"Claude Code 的高效持久化内存系统，在会话启动时可节省约 67% 的 token（经 tiktoken 验证）。分层架构，具有渐进式加载、紧凑编码、分支感知上下文、智能压缩、会话差异、冲突检测、会话续传协议和恢复模式。在会话开始时（如果存在 MEMORY.md）、在响应“记住这一点”、“继续我上次的进度”、“我们上次在做什么”、“总结”、“保存进度”、“别忘了”、“切换上下文”、“交接”、“内存健康”、“保存状态”、“继续我上次的进度”、“上下文预算”、“还剩多少上下文”或任何具有现有内存文件的项目会话开始时激活。此技能一次解决两个问题：Claude 在会话之间忘记一切，以及会话太快达到上下文限制。它用紧凑、结构化的内存加载取代了数千个浪费的重新解释 token，使 Claude 能够在 2000 个 token 内获得完整的项目上下文。\n",{"claudeCode":475},"Nagendhra-web/memory-bank","memory-bank","https://github.com/Nagendhra-web/memory-bank",{"basePath":479,"githubOwner":480,"githubRepo":476,"locale":18,"slug":476,"type":263},"skills/memory-bank","Nagendhra-web",{"evaluate":482,"extract":489},{"promptVersionExtension":213,"promptVersionScoring":214,"score":217,"tags":483,"targetMarket":268,"tier":225},[220,222,223,484,485,486,487,488],"sessions","token-efficiency","branch-aware","compression","ai-assistant",{"commitSha":270,"license":254},{"parentExtensionId":491,"repoId":492,"translatedFrom":493},"k17dbjbnpmrxhbszzw5sx6tsgs86n09d","kd740s537z66ds6qge39y9dk3186nbp9","k1723rcremvxvx59dy10cjc0f186n67k",[488,486,487,222,220,223,484,485],{"evaluatedAt":496,"extractAt":497,"updatedAt":498},1778694438899,1778694389360,1778694473202,{"_creationTime":500,"_id":501,"community":502,"display":503,"identity":509,"providers":514,"relations":522,"tags":525,"workflow":526},1778696691708.329,"k170yjw596k3cg892bq130jyd186mhyz",{"reviewCount":8},{"description":504,"installMethods":505,"name":507,"sourceUrl":508},"Validate a Claude Code plugin structure, frontmatter, and MCP tool references",{"claudeCode":506},"ruvnet/ruflo","Validate Plugin","https://github.com/ruvnet/ruflo",{"basePath":510,"githubOwner":511,"githubRepo":512,"locale":290,"slug":513,"type":263},"plugins/ruflo-plugin-creator/skills/validate-plugin","ruvnet","ruflo","validate-plugin",{"evaluate":515,"extract":520},{"promptVersionExtension":213,"promptVersionScoring":214,"score":516,"tags":517,"targetMarket":268,"tier":225},100,[518,318,519,224,428],"validation","developer-tools",{"commitSha":270,"license":521},"MIT",{"parentExtensionId":523,"repoId":524},"k17f4y1y2y777p7zrxxhbnf03n86mr5j","kd7ed28gj8n0y3msk5dzrp05zs86nqtc",[224,519,428,318,518],{"evaluatedAt":527,"extractAt":528,"updatedAt":527},1778701241052,1778696691708,{"_creationTime":530,"_id":531,"community":532,"display":533,"identity":539,"providers":543,"relations":552,"tags":555,"workflow":556},1778699349323.906,"k175ez1g485tsjr6ac09xr59hd86mmqk",{"reviewCount":8},{"description":534,"installMethods":535,"name":537,"sourceUrl":538},"当通过本地 LiteLLM 代理将 Claude Code 路由到 GitHub Copilot 时使用，以减少直接的 Anthropic 支出，配置 ANTHROPIC_BASE_URL 或 ANTHROPIC_MODEL 覆盖，或对 Copilot 代理设置失败进行故障排除，例如 model-not-found、无 localhost 流量或 GitHub 401/403 身份验证错误。",{"claudeCode":536},"xixu-me/skills","running-claude-code-via-litellm-copilot","https://github.com/xixu-me/skills",{"basePath":540,"githubOwner":541,"githubRepo":542,"locale":18,"slug":537,"type":263},"skills/running-claude-code-via-litellm-copilot","xixu-me","skills",{"evaluate":544,"extract":551},{"promptVersionExtension":213,"promptVersionScoring":214,"score":516,"tags":545,"targetMarket":268,"tier":225},[546,547,548,224,549,550],"copilot","litellm","proxy","configuration","troubleshooting",{"commitSha":270},{"repoId":553,"translatedFrom":554},"kd77r2vb42jmgam0qbr9f2c6kn86mebv","k17c1n6dbde7qh2wemx78d13t986n7x2",[224,549,546,547,548,550],{"evaluatedAt":557,"extractAt":558,"updatedAt":559},1778699196029,1778699106670,1778699349323,{"_creationTime":561,"_id":562,"community":563,"display":564,"identity":570,"providers":574,"relations":582,"tags":586,"workflow":587},1778696876879.2397,"k178m626tpcxa8b3bekz8y0x9n86m9w1",{"reviewCount":8},{"description":565,"installMethods":566,"name":568,"sourceUrl":569},"保护 Claude Code 会话免受上下文溢出影响，通过运行一个后台守护进程来监控会话大小并在压缩命中之前自动进行修剪。当用户说“guard”、“protect session”、“context getting long”、“prevent compaction”、“session management”或正在运行需要持续上下文保护的代理团队时使用。\n",{"claudeCode":567},"Ruya-AI/cozempic","guard","https://github.com/Ruya-AI/cozempic",{"basePath":571,"githubOwner":572,"githubRepo":573,"locale":18,"slug":568,"type":263},"plugin/skills/guard","Ruya-AI","cozempic",{"evaluate":575,"extract":581},{"promptVersionExtension":213,"promptVersionScoring":214,"score":516,"tags":576,"targetMarket":268,"tier":225},[577,578,224,579,580],"session-management","context-pruning","python","utility",{"commitSha":270},{"parentExtensionId":583,"repoId":584,"translatedFrom":585},"k176hd1j2vn0hpak7ds6v3eand86mfqh","kd79d77qmyh3826dwhk7ynx9xd86nmqm","k17ef0e0f693nd248cs3f63rys86ndy8",[224,578,579,577,580],{"evaluatedAt":588,"extractAt":589,"updatedAt":590},1778696790816,1778696712852,1778696876879,{"_creationTime":592,"_id":593,"community":594,"display":595,"identity":601,"providers":605,"relations":612,"tags":615,"workflow":616},1778694480889.9487,"k17eg219cke0p45yfhh8n2t37986mgm4",{"reviewCount":8},{"description":596,"installMethods":597,"name":599,"sourceUrl":600},"Interactive assistant for creating new Claude commands with proper structure, patterns, and MCP tool integration",{"claudeCode":598},"NeoLabHQ/context-engineering-kit","create-command","https://github.com/NeoLabHQ/context-engineering-kit",{"basePath":602,"githubOwner":603,"githubRepo":604,"locale":290,"slug":599,"type":263},"plugins/customaize-agent/skills/create-command","NeoLabHQ","context-engineering-kit",{"evaluate":606,"extract":611},{"promptVersionExtension":213,"promptVersionScoring":214,"score":516,"tags":607,"targetMarket":268,"tier":225},[608,609,610,428,224],"command-creation","agent-development","prompt-engineering",{"commitSha":270},{"parentExtensionId":613,"repoId":614},"k173g0c46wghgz5j9gn02wfydn86mkh8","kd7a3rj13ezgx1wgm0jfh08hsx86n0sz",[609,224,608,428,610],{"evaluatedAt":617,"extractAt":618,"updatedAt":617},1778694773368,1778694480890,{"_creationTime":620,"_id":621,"community":622,"display":623,"identity":629,"providers":634,"relations":641,"tags":644,"workflow":645},1778692488329.0164,"k17d7dgccv6vnj9tcw0cehwnph86nryb",{"reviewCount":8},{"description":624,"installMethods":625,"name":627,"sourceUrl":628},"Analyze which rules are actively used vs inert. Detect coverage gaps. Recommend pruning to reduce token consumption.",{"claudeCode":626},"luiseiman/claude-kit","Rule Effectiveness Analysis","https://github.com/luiseiman/claude-kit",{"basePath":630,"githubOwner":631,"githubRepo":632,"locale":290,"slug":633,"type":263},"skills/rule-effectiveness","luiseiman","claude-kit","rule-effectiveness",{"evaluate":635,"extract":640},{"promptVersionExtension":213,"promptVersionScoring":214,"score":516,"tags":636,"targetMarket":268,"tier":225},[224,549,637,638,639],"analysis","optimization","git",{"commitSha":270,"license":521},{"parentExtensionId":642,"repoId":643},"k17650xadq8363kzchr4w0gmf186nq24","kd79wqc8an5wh20cc2znr8tyb586mxwx",[637,224,549,639,638],{"evaluatedAt":646,"extractAt":647,"updatedAt":646},1778693275143,1778692488329]