PiSkillPiSkill
#ai knowledge base#knowledge graph#agent memory#persistent memory#document ingestion

Mini Context Graph

Independent PiSkill directory guide. The original skill remains hosted by GitHub Awesome Copilot.

What is Mini Context Graph?

Creates a persistent knowledge base and lightweight knowledge graph that ingests documents, writes linked wiki-style pages, extracts entities and relations, and keeps source material for evidence retrieval.

What does Mini Context Graph do?

Mini Context Graph is a GitHub Awesome Copilot skill for building persistent agent knowledge from documents using three linked layers: immutable source material, maintained wiki pages and a knowledge graph of entities and relationships. The goal is to compound knowledge across sessions instead of re-deriving the same context on every query.

Who is Mini Context Graph best for?

  • Teams building persistent AI knowledge bases
  • Agent systems that need memory across sessions
  • Projects combining document evidence with structured relationships
  • Developers exploring alternatives to query-only RAG

Common use cases

  • Ingest documents into a persistent knowledge base
  • Maintain wiki-style summaries and entity pages
  • Extract entities and relationships into a graph
  • Trace synthesized knowledge back to raw source evidence

How does Mini Context Graph work?

The skill stores original documents as evidence, has the LLM create and update wiki pages, and persists extracted entities and relationships in a graph. The wiki and graph are intended to grow over time while provenance links retain the path back to source text.

Key benefits

  • Combines narrative knowledge and structured relationships
  • Avoids rebuilding all context from scratch on every query
  • Keeps raw evidence available for provenance
  • Supports structural graph traversal as well as written summaries

Things to know

  • It introduces persistent storage and maintenance complexity compared with simple RAG
  • Entity extraction and summaries can still contain model errors and need validation
  • Its implementation is specific to the repository's supplied scripts and data structures

Compatible tools

GitHub Copilot

Frequently asked questions

How can I build an AI knowledge base with persistent context?
Mini Context Graph combines stored source documents, maintained wiki pages and a persistent entity-relation graph so knowledge can accumulate across sessions.
How is this different from standard RAG?
The source skill explicitly contrasts itself with re-discovering context for every query: it persists summaries, entities and relations while retaining the original source evidence.

Related skills

Agent Systems

Pinecone RAG

Builds production RAG pipelines and persistent agent memory with Pinecone, covering document indexing, embeddings, hybrid retrieval, searchable knowledge bases, namespaces, and multi-session memory.

Agent Systems

Remember / Agent Memory

Turns lessons from debugging and completed work into organized, reusable memory instructions at user or workspace scope so future AI sessions can reuse hard-won project knowledge.

Agent Systems

AI Team Orchestration

Coordinates AI-assisted software work across planning, implementation, QA, review, durable project context, handoffs, and recovery from interrupted or multi-session work.

Agent Systems

GitHub Copilot SDK

Helps developers embed GitHub Copilot's agent runtime into applications with Python, TypeScript, Go, or .NET, including tools, sessions, streaming, MCP, and custom agents.

Agent Systems

Structured Autonomy Implementation

Executes an approved implementation plan step by step, keeps progress state inside the plan, runs specified checks, and stops at explicit handoff points rather than improvising outside scope.

Agent Systems

Suggest Relevant Copilot Skills

Scans the current repository and available GitHub Copilot skills, then recommends relevant skills, avoids duplicates, and identifies installed skills that may be outdated.

← Back to Skills Directory