cobol-kg

name	cobol-kg
description	COBOL Knowledge Graph analysis and debugging toolkit. Use this skill when working with Neo4j queries, debugging pipeline issues, analyzing COBOL programs, or running common development tasks for this project.

COBOL Knowledge Graph Skill

This skill provides commands and workflows for the COBOL Agentic Knowledge Graph project.

Project Overview

This project builds a knowledge graph from legacy COBOL codebases using:

• Neo4j for graph storage
• LangChain/LangGraph for agent orchestration
• Multiple LLM providers (OpenAI, Groq, Gemini)

Quick Reference Commands

Check Neo4j Connection & Counts

uv run python -c "
from cobol_agentic_kg.utils.neo4j_client import Neo4jClient
client = Neo4jClient()
result = client.run_query('MATCH (n) RETURN labels(n)[0] as label, count(*) as count ORDER BY count DESC')
for r in result: print(f'{r[\"label\"]}: {r[\"count\"]}')
client.close()
"

Common Cypher Queries

List all programs:

MATCH (p:Program) RETURN p.name, p.type, p.complexity ORDER BY p.name

Find program dependencies:

MATCH (p:Program)-[:CALLS]->(called:Program)
WHERE p.name = 'PROGRAM_NAME'
RETURN p.name, called.name

Find data flow:

MATCH (p:Program)-[:USES]->(d:DataItem)
WHERE p.name = 'PROGRAM_NAME'
RETURN d.name, d.type, d.level

Find JCL jobs for a program:

MATCH (j:Job)-[:EXECUTES]->(p:Program)
WHERE p.name = 'PROGRAM_NAME'
RETURN j.name, j.description

Find copybooks used: