Hacker
Mohammed Ali Khan
@alikhan207
Traceback: understand the code, trace the decisions, find the people who know
Graph HacksOct 2026
When you join a software project, three questions slow you down: who actually knows this file, what could break if I change it, and why was it built this way? Code search can't answer them, because the answers live in the history and the people behind the code. Traceback is an agent that answers them from a FalkorDB knowledge graph of a real open-source project (pallets/flask): source files, imports, people, commits, pull requests and issues. It has no code search and no memorized Flask knowledge. It can only answer by calling graph tools. Each answer shows a plain-English summary, cards for the people, the files that could be affected and a timeline of past changes linking to the real GitHub pull requests and issues, and a drawn graph of the exact path it followed. It is for new contributors, maintainers doing impact analysis before a change, and engineering managers who need to know who to ask. Flask is our test dataset. Expertise ranking is a proxy based on commits, PRs and reviews, and we say so in the app.