clawhub-mem-research

专门用于从 ClawHub.ai 抓取、筛选并调研与“记忆（Memory/Mem）”相关的最新 OpenClaw Skill。该 Skill 能够自动识别指定日期的更新，判断 Skill 与记忆功能的相关性，并生成结构化的技术调研报告。适用场景：(1) 追踪 ClawHub 上的记忆技术趋势，(2) 自动化生成特...

Security Scan

The skill's requirements and runtime instructions are coherent with its stated purpose: it only instructs the agent to browse ClawHub search results, read SKILL/README pages of up to 3 items, and produce/save a Markdown report — no unrelated credentials, installs, or privileged actions are requested.

Purpose & Capability

Name and description claim to scrape and research memory-related OpenClaw skills; the SKILL.md instructs exactly that (visit a clawhub.ai search URL, inspect top 3 results, read their SKILL.md/README, and summarize). No unrelated env vars, binaries, or installs are requested.

Instruction Scope

Instructions are narrowly scoped to visiting the specified ClawHub search URL, filtering the first 3 results for relevance, reading SKILL.md/README of each, extracting technical details, and writing a Markdown report to report_v1/YYYY-MM-DD/clawhub-mem-research.md. Minor note: it directs the agent to read arbitrary third-party SKILL.md/README files (expected for this purpose) — those files may contain any content, links, or code snippets; the skill does not instruct executing any code found, only reading and summarizing.

Install Mechanism

No install spec and no code files are present — instruction-only skill. This is the lowest-risk install surface and consistent with the described functionality.

Credentials

The skill requests no environment variables, credentials, or config paths. That is proportionate for a web-scraping / research task that targets a public site.

Persistence & Privilege

always is false and the skill does not request persistent system-wide changes or modify other skills. It does direct the agent to write an output file to a local directory (report_v1/YYYY-MM-DD), which is expected for a reporting tool.

Assessment

This skill appears coherent and limited: it will browse the specified ClawHub search page, read up to 3 skills' SKILL.md/README content, produce a Chinese Markdown report, and save it to a local report_v1/YYYY-MM-DD directory. Before enabling or running it, confirm the agent has (1) a web-browsing capability to access clawhub.ai, and (2) permission to write files to the agent runtime filesystem. Be aware that the skill will read arbitrary third-party SKILL.md/README contents (which may contain links or pasted content); it does not request credentials or ask to execute code, but you should ensure the agent does not follow or execute third-party links unless you trust them. If you need stricter limits, require manual approval before the agent fetches external pages or change the save location/filename policy.

ClawHub 记忆技术调研工具 (ClawHub Mem Research)

本 Skill 用于自动化调研 ClawHub 上与记忆相关的最新技术。

工作流程

1. 访问与搜索： 使用浏览器访问 https://clawhub.ai/skills?sort=updated&dir=desc&nonSuspicious=true&q=memory。

2. 筛选准则：

   • 数量限制：仅读取前 3 条结果。
   • 相关性判断：通过标题和摘要判断该 Skill 是否真正属于“记忆增强/管理”范畴（如：存储事实、遗忘曲线、向量检索、内存压缩等）。如果只是名字包含 mem 但功能无关（如 Member 管理），则跳过。

3. 深度调研： 对于符合条件的每一个 Skill，点击进入详情页，重点阅读 SKILL.md 或 README.md，提取核心功能和技术实现细节。

4. 报告生成： 按照以下固定格式为每个技术生成摘要，并汇总成 Markdown 文件。

报告格式

每个技术项应包含以下详细结构，内容应根据 Skill 的实际技术特性灵活组织：

[序号]. [技术名称]

—— [一句话副标题/定位描述]

• 核心功能： 深入描述该技术的核心作用、解决的痛点以及应用场景（如：模拟 REM 睡眠、遗忘曲线管理等）。

• 技术要点：

  • 根据 Skill 的实际实现，提取其最关键的技术特征。
  • 采用分点/小标题形式（如：动态强度衰减算法、本地向量化检索等）。
  • 详细解释实现逻辑、算法公式、触发机制、存储方式或控制指令。
  • 不要生搬硬套固定模板，应以最能体现该技术价值的方式展示其技术细节。

• SKILL链接：技术名称 (版本号)

输出要求

• 语言：必须使用中文编写报告。
• 存储：调研完成后，将生成的报告保存到 report_v1/YYYY-MM-DD 目录下，文件命名格式为 clawhub-mem-research.md（YYYY-MM-DD 为调研的目标更新日期）。