Token导航 LogoToken导航TokenDH.com
研究检索external-servicegithub未标认证来源可访问clear审计通过

memory记忆

Agent Skill

memory 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

321

周安装

13

GitHub Stars

9

下载量

101
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

3

许可证

MIT

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:memory(记忆)
来源仓库:https://github.com/tao3k/omni-dev-fusion
仓库路径:skills/memory
安装命令:
npx skills add https://github.com/tao3k/omni-dev-fusion --skill memory
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/tao3k/omni-dev-fusion --skill memory

简介

memory 用于查找、检索和筛选相关信息,支持多源信息整合。

  • 适用于关键词搜索、任务场景匹配或来源线索梳理等研究检索场景。
  • 通过关键词输入和来源仓库路径获取候选信息,支持进一步核验具体用法。
  • 安装前需确认权限范围、维护状态,以及是否会触发联网或文件读写操作。
  • memory 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Memory Skill Policy

This skill is an MCP-facing facade for memory operations. Core memory policy (lifecycle/revalidation/promotion) belongs to Rust memory core, not this skill surface.

Router Logic

Scenario 1: User wants to store a transient finding

  1. Analyze: Determine if the item is transient operational memory (bug/workaround/incident note)
  2. Store: Call save_memory(content, metadata)
  3. Confirm: Show the saved memory ID

Scenario 2: User wants to remember/search

  1. Search: Call search_memory(query, limit)
  2. Format: Present results with relevance scores
  3. Respond: "I found X memories about that..."

Scenario 3: User asks for current operational memory status

  1. List: Call get_memory_stats()
  2. Recall: Call search_memory() with relevant keywords
  3. Present: Show structured summary with transient scope

Commands Reference

CommandDescriptionExample
save_memoryStore short-term operational memory itemsave_memory("Temporary workaround for timeout in parser", {"tag": "incident"})
search_memorySemantic search in memorysearch_memory("git commit format", limit=5)
index_memoryOptimize vector index (IVF-FLAT)index_memory()
get_memory_statsGet memory countget_memory_stats()
load_skillLoad skill manifest into memoryload_skill("git")

Workflow: Store a Transient Workaround

User: Remember this temporary fix: increase parser timeout when MCP queue spikes.

Claude:
  1. save_memory(
       content="Temporary workaround: increase parser timeout under MCP queue spike",
       metadata={"domain": "runtime", "kind": "workaround", "source": "user"}
     )
  2. → Saved memory [a1b2c3d4]: Temporary workaround: increase parser timeout...
  3. → "Stored as short-term operational memory."

Workflow: Recall Recent Operational Context

User: What workaround did we use for MCP queue timeout?

Claude:
  1. search_memory("MCP queue timeout workaround")
  2. → Found 2 matches:
     - [Score: 0.8921] Temporary workaround: increase parser timeout...
     - [Score: 0.7234] Queue backpressure note...
  3. → "I found recent operational memory for this issue..."

Memory vs Knowledge Skill

AspectMemory (this skill)Knowledge skill
ScopeShort-term operational contextLong-term reusable knowledge
NatureTransient (can be purged after revalidation)Stable (promoted/curated)
Purpose"What recent issue/workaround context exists?""What durable rule/pattern should be reused?"
PolicyManaged by Rust memory core lifecycleManaged by knowledge ingestion/curation flows
ExposureMCP tool facadeMCP tool interface

Best Practices

  1. Store transient operational memory, not permanent rules
  2. Include kind in metadata (incident, workaround, observation)
  3. Use clear, searchable phrasing in content
  4. Promote only proven durable patterns to knowledge

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

04

需要参考平台分布和安装热度时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Claude Code

29.3%
按下载量换算30

windsurf

21.13%
按下载量换算21

trae

16.8%
按下载量换算17

OpenCode

12.2%
按下载量换算12

Codex

6.7%
按下载量换算7

Antigravity

3.1%
按下载量换算3

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

external-service

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。

来源信息

继续浏览同类 Skills