返回 Skill 列表
extension
分类: AI Agent 能力无需 API Key

Serena Memory System

长期记忆和上下文管理接口。用于存储和检索用户偏好、项目上下文以及过去的学习内容。

person作者: jakexiaohubgithub

Serena Memory System Skill

This skill grants access to the serena MCP tools, which provide persistent memory and context awareness for the agent.

When to use

  • Recalling user preferences or project history.
  • Storing new "memories" or learnings for future sessions.
  • searching for files or symbols using Serena's semantic index.

Available Tools (Context Loaded)

The following tools are available via the serena MCP server:

Memory Management

  • mcp__serena__list_memories: Retrieve stored memories.
  • mcp__serena__think_about_collected_information: Process and synthesize information.
  • mcp__serena__onboarding: Access onboarding information.

Semantic Search

  • mcp__serena__find_file: Find files semantically.
  • mcp__serena__find_symbol: Find code symbols semantically.

Best Practices

  1. Check Memory First: Before asking the user for repetitive info, check if it's stored in memory.
  2. Synthesize: Use think_about_collected_information to consolidate complex context.

Example Workflow

  1. User: "What was the design pattern we decided on?"
  2. Agent: Calls mcp__serena__list_memories to search for design decisions.