skills.rest
senoldogann
senoldogannCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

context-manager

Transform static source code into a queryable Knowledge Graph for AI agents.

62|3|Updated Jan 8, 2026
Search Tags:#knowledge graph#codebase#semantic search#mcp#graph traversal#ast parsing
One-click install
npx skills add https://github.com/senoldogann/LLM-Context-Manager --skill context-manager-senoldogann
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: context-manager
Source: https://github.com/senoldogann/LLM-Context-Manager/tree/main
Command: npx skills add https://github.com/senoldogann/LLM-Context-Manager --skill context-manager-senoldogann

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

AI coding assistants suffer from critical blindness: they cannot see entire large codebases, hallucinate dependencies without structural understanding, and lose context when using simple keyword or vector search. This Skill transforms static source code into a dynamic, queryable Knowledge Graph, enabling AI agents to navigate, understand, and reason about any project with surgical precision.

Core Features & Use Cases

  • Semantic Hybrid Search: Combines vector similarity and graph centrality to find code by meaning, not just keywords, with confidence scores and explanations.
  • Graph Traversal & Impact Analysis: Trace call chains between functions, find all callers of a node, and assess the blast radius of changes before editing.
  • Cursor-Aware Context Retrieval: Get relevant surrounding code and graph neighbors at any specific file and line position.
  • Multi-Language Support: Parses Rust, Python, TypeScript, JavaScript, Go, Java, Kotlin, and C# using Tree-sitter AST, plus data files like Markdown, JSON, and YAML.

Quick Start

Install the context-manager skill, configure your MCP client with the one-line installer, and ask your AI agent to index your project and search for code relationships using natural language queries.

Frequently Asked Questions about context-manager

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I perform semantic code search and call chain tracing across a large codebase?▼

To perform semantic search and call chain tracing across a large codebase, this Skill transforms source code into a queryable knowledge graph using AST parsing and vector search. It enables AI agents to find code by meaning and trace function callers with surgical precision.

Can I analyze blast radius and code impact for Rust, Python, and TypeScript projects?▼

Yes, you can analyze blast radius and code impact for Rust, Python, TypeScript, JavaScript, Go, Java, Kotlin, and C# projects. The Skill uses Tree-sitter AST parsing and graph traversal to assess the structural impact of changes before editing.

What is the best way to retrieve cursor-aware context for an AI coding assistant?▼

The best way to retrieve cursor-aware context is by transforming source code into a knowledge graph that combines vector similarity and graph centrality. This provides AI agents with relevant surrounding code and graph neighbors at any specific file and line position.

How does graph traversal improve codebase intelligence compared to keyword search?▼

Graph traversal improves codebase intelligence by mapping structural relationships rather than relying on keyword matching. It combines LanceDB vector search with Petgraph graph traversal to trace call chains, find callers, and calculate confidence scores with explanations.

Do I need an MCP client to index my project and query code relationships?▼

Yes, you need an MCP client to use the one-line installer and query code relationships. Once configured, you can ask your AI agent to index your project and search for code relationships using natural language queries.

Why do AI coding assistants hallucinate dependencies in large codebases?▼

AI coding assistants hallucinate dependencies because they lack structural understanding of large codebases and lose context with simple keyword search. Transforming static source code into a dynamic knowledge graph solves this by providing graph-based context retrieval.

Related Skills

Explore More Software Engineering Skills→
Alex1980AlexAlex1980Alex

pdf-knowledge

Index PDFs into vector and graph knowledge stores for search and extraction.

Community
Advanced
tractorjuicetractorjuice
2.1k

arckit-impact

Traverse reverse dependencies to classify impact severity and recommend remediation commands.

Community
Advanced
davpatel605-beepdavpatel605-beep

qdrant-vector-search

Build vector similarity search systems with Qdrant for RAG applications.

Community
Advanced
aws-samplesaws-samples
36

searching

Retrieve cited answers from documents and knowledge graphs with hybrid search.

Official
Intermediate