← 返回 Skills 市场
ClawBrain Recall
作者
michaelfeng
· GitHub ↗
· v1.2.0
· MIT-0
216
总下载
0
收藏
0
当前安装
8
版本数
在 OpenClaw 中安装
/install clawbrain-pro-recall
功能描述
v1.2 精准回忆能力:关键词与语义双重搜索、来源标注、相关度排序。精准找回历史对话和决策。
使用说明 (SKILL.md)
ClawBrain Recall
让你的龙虾拥有"回忆"能力——通过知识图谱精准检索过去的对话、决策和关键信息。
解决什么问题
默认的 OpenClaw Agent 只能看到当前对话上下文。一旦对话结束或上下文被压缩,之前的内容就"忘了"。Recall 让 Agent 能够:
- 搜索历史对话中的实体和关系
- 查找之前的决策和讨论结论
- 检索关联信息(提到A,自动找到相关的B和C)
- 基于历史上下文给出更准确的回答
v1.2 核心能力
搜索精准度
搜索采用双重匹配策略:
- 精确关键词搜索:输入确切的名称、术语,直接命中
- 语义搜索:用自然语言描述你想找的内容,即使措辞不同也能匹配
搜索结果会显示相关度百分比,帮你快速判断哪条回忆最相关。
来源标注
每条回忆结果都会标注来源类型:
- 原话:你在过去对话中说过的原始表述
- 摘要:AI 根据多次对话归纳出的总结
让你清楚地知道信息的可靠程度。
降级透明
当搜索服务暂时不可用时,Agent 会主动通知你"当前无法搜索记忆",并建议替代方案,而不是默默返回空结果。
触发方式
以下表达会自动触发回忆搜索:
- "之前我们讨论过..."
- "你还记得...吗"
- "找一下上次的..."
- "之前那个决定是什么来着"
建议配合使用
- clawbrain-pro-memory:主动记录 + 被动回忆,完整的记忆体系
- ClawBrain Auto:智能决策引擎,复杂问题深度思考
了解更多:https://clawbrain.dev
安全使用建议
This is a descriptive, instruction-only skill that doesn't itself store or access memories or secrets. Before relying on it, confirm which concrete memory backend (e.g., 'clawbrain-pro-memory' or another storage service) your agent will use and review that backend's permissions and data-handling practices. If you care about privacy, verify where historical conversations are stored and who can access them.
功能分析
Type: OpenClaw Skill
Name: clawbrain-pro-recall
Version: 1.2.0
The skill bundle (clawbrain-pro-recall) provides documentation and metadata for a conversation retrieval feature. The SKILL.md file describes standard RAG (Retrieval-Augmented Generation) capabilities such as keyword/semantic search and source tagging for historical dialogue. No executable code, malicious instructions, or data exfiltration triggers were identified in the provided files.
能力评估
Purpose & Capability
The name and description claim a conversational 'recall' capability and the SKILL.md contains only high-level instructions and expected behaviors for retrieving past dialogue; nothing requested (no env vars, binaries, or config paths) is inconsistent with that purpose.
Instruction Scope
SKILL.md is prose-only and does not instruct the agent to read files, access environment variables, call external endpoints, or execute commands. It lists trigger phrases for invoking recall, which is consistent with its stated purpose.
Install Mechanism
No install spec and no code files are present (instruction-only), so nothing is written to disk or fetched during install.
Credentials
The skill declares no required environment variables, credentials, or config paths — appropriate for a description-only recall helper.
Persistence & Privilege
always is false and the skill is user-invocable; there is no request for permanent presence or modifications to other skills/configuration.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install clawbrain-pro-recall - 安装完成后,直接呼叫该 Skill 的名称或使用
/clawbrain-pro-recall触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.2.0
精准搜索+相关度百分比+来源标注
v1.0.2
Fix display name
v1.0.1
v1.0.1: 重写为图谱检索 (N-gram + 2-hop 遍历 + API 端点)
v0.9.5
优化 summary 文案
v0.9.4
统一品牌名
v0.9.3
更新 summary 描述
v0.9.2
v0.9.2: 记忆检索优化
v1.0.0
- Initial release of clawbrain-pro-recall.
- Adds "recall" capabilities to search previous conversations, locate old files, and retrieve archived memories.
- Enables agents to reference past context, overcoming the default memory limitations.
- Searches recent and archived memory files, as well as workspace data, to provide accurate responses based on history.
- Triggered automatically by user prompts referring to past discussions or files.
- Integrates well with related skills for a complete memory system.
元数据
常见问题
ClawBrain Recall 是什么?
v1.2 精准回忆能力:关键词与语义双重搜索、来源标注、相关度排序。精准找回历史对话和决策。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 216 次。
如何安装 ClawBrain Recall?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install clawbrain-pro-recall」即可一键安装,无需额外配置。
ClawBrain Recall 是免费的吗?
是的,ClawBrain Recall 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
ClawBrain Recall 支持哪些平台?
ClawBrain Recall 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 ClawBrain Recall?
由 michaelfeng(@michaelfeng)开发并维护,当前版本 v1.2.0。
推荐 Skills