Beyond Vibe Coding: AI Agents for Technical Debt Detection
summary
An extensible framework uses Dagger to spin up ephemeral containers and Neo4j to store a code graph built from Python, JavaScript and TypeScript parsers; the graph contains file and symbol nodes plus DEFINED_IN, IMPORTS, CALLS, CONTAINS and re-export edges. Cypher queries run via a cypher-shell client container as CI jobs, detectors are simple Cypher rules (generated with LLM prompts but executed without tokens), and a proof‑of‑concept agent generates missing unit tests from the graph; demo is prerecorded (15 minutes) due to dependency bugs.
AI-written summary · about
chapters (19)
- 0:00Introduction and setup
- 1:55What is Dagger and why use it
- 4:45Installing Dagger and Neo4j
- 6:31Exploring Dagger traces
- 7:56Neo4j service in Dagger
- 9:03Visualizing code structure
- 10:06Dagger module structure
- 11:49Code smell detection workflow
- 13:13Integrating with CI
- 16:06Q&A session
- 16:51AI agents and Neo4j integration
- 18:56Technical debt detection workflows
- 21:11Agentic RAG with Neo4j
- 22:37Adding graph to MCP in IDE
- 23:22Open source repositories
- 24:47Ephemeral services and flexibility
- 25:41Dagger's evolving tooling
- 26:47Agentic containers in Kubernetes
- 27:30Closing remarks and Q&A
related talks
38mSneak peek: Vibe coding AI agents course
33mWhen Vibe Coding Meets Agent Engineering: The Next Evolution of Software Architecture
27mFantastic Vibes and Where To Find Them
48mAI Coding Agents and How to Code Them
47mExploring a complex codebase with AI
30mBeyond the Hype: Using AI Coding Agents Today
42mBeyond the Vibe, How far can we push It?
40mAI Coding Agents and how to code them