engram-distill

Name: engram-distill
Availability: InStock
Author: PhaniJ

Distill unstructured documentation and code into a queryable knowledge graph via MCP tool calls.

Updated Jun 29, 2026

Search Tags:#ontology #mcp #bridging #knowledge-graph #code-understanding #agentic-memory

Ready to run

npx skills add https://github.com/phanijapps/engram --skill engram-distill

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: engram-distill
Source: https://github.com/phanijapps/engram/tree/main/.claude/skills/engram-distill
Command: npx skills add https://github.com/phanijapps/engram --skill engram-distill

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes assets (resource) components.

What problem does it solve?

This skill solves the problem of fragmented, disposable AI context by enabling the creation of a persistent, source-grounded knowledge graph that links conceptual understanding directly to code and documentation.

Core Features & Use Cases

Semantic Bridging: Connects high-level domain concepts to specific code symbols and documentation artifacts, preventing orphan knowledge.
Multi-Layer Knowledge Graph: Organizes information into technical, documentation, and conceptual layers for reliable, auditable recall.
Use Case: When onboarding to a complex codebase, use this skill to distill RFCs and design notes into the knowledge graph, ensuring the agent understands the business logic behind the implementation.

Quick Start

Use the engram-distill skill to scan the current repository and bridge the authentication concepts found in the documentation to the corresponding middleware code.

Frequently Asked Questions about engram-distill

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

FAQPage Schema

How do I build a queryable knowledge graph from unstructured documentation and code?▼

To build a queryable knowledge graph from unstructured documentation and code, distill the source materials into entities and relationships using deterministic MCP tool calls. This process organizes information into technical, documentation, and conceptual layers for reliable recall.

How do I connect high-level domain concepts to specific code symbols during onboarding?▼

To connect high-level domain concepts to specific code symbols, use semantic bridging within a project-scoped workspace. This links domain understanding directly to technical artifacts and documentation, preventing orphan knowledge and ensuring agents grasp business logic.

Do I need a configured MCP server to persist agentic memory for codebase understanding?▼

Yes, you need a configured engram-mcp server to persist agentic memory and maintain idempotency across extraction batches. The server manages the knowledge graph storage and ensures semantic bridges remain auditable during codebase understanding.

What is the best way to structure durable agentic memory for long-running AI projects?▼

The best way to structure durable agentic memory is to organize extracted knowledge into multi-layer graphs separating technical, documentation, and conceptual data. This structured approach ensures persistent, source-grounded recall across long-running AI project sessions.

Why does my AI agent lose context about business logic when navigating complex codebases?▼

AI agents lose context about business logic because standard sessions use disposable memory. By distilling RFCs and design notes into a persistent knowledge graph, the agent maintains a structured, auditable link between conceptual domain logic and implementation artifacts.

Can I manage knowledge graph entities and relationships within a project-scoped workspace?▼

Yes, you can manage knowledge graph entities and relationships within a project-scoped workspace. The workspace isolates extraction batches and semantic bridges, ensuring deterministic MCP tool calls maintain data integrity and idempotency for the specific project.