encode-repo-serena

rjmurillo 的头像作者 rjmurillo

使用Serena的LSP驱动的符号分析系统地填充遗忘知识库,以实现准确、全面的理解代码库。

📖 中文导读:使用Serena的LSP驱动的符号分析系统地填充遗忘知识库,以实现准确、全面的理解代码库。 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。

name: encode-repo-serena version: 1.0.0 description: Systematically populate the Forgetful knowledge base using Serena's LSP-powered symbol analysis for accurate, comprehensive codebase understanding. license: MIT model: claude-sonnet-4-5 metadata: argument-hint: 'Project path or name to encode (default: current directory)'

Encode Repository (Serena-Enhanced)

Transform an undocumented codebase into a rich, searchable knowledge repository using Serena's LSP-powered symbol analysis.

Quick Start

/encode-repo-serena
/encode-repo-serena ./my-project
"encode this repository"
"populate forgetful with this codebase"
Input Output Duration
Codebase path Forgetful memories + entities + docs 30-60 min

Prerequisites

  1. Serena plugin: claude plugins list | grep serena
  2. Forgetful MCP: Test with execute_forgetful_tool("list_projects", {})
  3. If missing, run /context-hub-setup first

Phase Overview

Phase Focus Output
0 Discovery Project assessment, structure map
1 Foundation 5-10 project overview memories
1B Dependencies 1-3 dependency memories
2 Symbols 10-15 architecture memories
2B Entities Component entities + relationships
3 Patterns 8-12 pattern memories
4 Features 1-2 per critical feature
5 Decisions Design decision memories
6 Artifacts Code artifact storage
6B Symbol Index Document + entry memory
7 Documents Long-form documentation
7B Architecture Architecture reference doc

See references/phases.md for full phase details.

Memory Targets

Profile Total Memories Documents Entities
Small Simple 17-31 2 3-5
Small Complex 28-46 2 5-10
Medium 38-66 2-3 10-20
Large 66-112 3-6 20-40

Execution Order

0 → 1 → 1B → 2 → 2B → 3 → 4 → 5 → 6 → 6B → 7 → 7B

Guidelines:

  • Execute phases in order
  • Use Serena's find_symbol and find_referencing_symbols
  • Deduplicate entities before creating
  • Link entities to memories bidirectionally
  • Create entry memories for documents

Quality Principles

Principle Description
Symbol-accurate Use LSP data, not guesses
Atomic One concept per memory
Size 200-400 words ideal
Importance Most should be 7-8
Linking Connect related memories

Validation Checklist

After completion:

  • Test memory search: "How do I add a new API endpoint?"
  • Test dependency query: "What dependencies does this project use?"
  • List entities by project
  • Verify entity relationships
  • Check Symbol Index document exists
  • Check Architecture Reference document exists
  • Verify project.notes populated

See references/validation.md for test commands.

References

Document Content
phases.md Detailed phase workflows
templates.md Entity schemas, memory templates
validation.md Validation test commands
  • /context-hub-setup - Setup Forgetful MCP
  • /using-forgetful-memory - Memory best practices
  • /using-serena-symbols - Serena symbol analysis