AI product Open source
Agent Context Workshop is a hands-on workshop resource for building and evaluating context layers for coding agents. It uses selected crates from the Qdrant Rust codebase to extract syntax with tree-sitter, resolve symbols and relationships with rustdoc JSON, store the resulting code knowledge graph in Neo4j, and expose graph tools such as search, blast-radius analysis, and implementor lookup to an agent.
The workshop compares graph-based agents with grep-style agents using gold-verified questions, deterministic scoring, cached evaluation results, tool-adoption measurements, and trace inspection. Additional notebooks cover static Python resolution with Griffe and an optional Fenic-and-LLM semantic layer. The GitHub repository maintains the workshop through ordered Jupyter notebooks, typed extraction and graph code, evaluation harnesses, tests, committed Qdrant seed data, and a committed evaluation cache; it is licensed under Apache-2.0.
1 use taken from transcripts — each links to the moment in the video.
Brandon Callender's GitHub repository for the hands-on workshop. It contains Jupyter notebooks that build a code context graph over a sample of the Qdrant Rust codebase (using tree-sitter, rustdoc/JSON symbol resolution, and Neo4j), a blast-radius tool wired to a Pydantic agent, and a second notebook comparing graph-tool agents against grep-style agents with Pydantic Evals, including pass@k metrics, tool-adoption measurement, and cached eval results. A third notebook covers applying the approach to Python. Dependencies include Docker for Neo4j and optionally an OpenRouter key.
1 in the library.