engyossefyossry-crypto/autonomous-devsecops-triage-engine

This project is an automated incident response system built on the n8n workflow platform. When a production error, crash, or stack trace comes in, either through a webhook or a scheduled morning check, the system uses a team of AI models working in sequence to investigate the problem, judge how serious it is, and notify the right people, instead of a person having to do all of that manually. The workflow is organized as a supervisor directing three specialized helper agents. One model acts purely as a router, keeping the process on track. A second model, chosen for handling long text well, acts like a forensic analyst, tracing the raw error logs back to the specific broken file and line of code and suggesting a fix. A third model reviews that finding from a compliance and security angle, checking things like whether an API key was exposed or whether the issue could involve data privacy rules, and rating the risk as critical, high, medium, or low. A fourth model then turns all of that technical detail into two kinds of messages: a detailed technical alert for engineers and a calmer, plain-language update for non-technical stakeholders. Each step in the process writes its findings into n8n's built in data tables, so there is a running record of debugging attempts, risk evaluations, and notification status that survives even if a later step, like sending a Slack message, happens to fail. The finished results are then sent out to multiple destinations at once, including a GitHub issue, a shared spreadsheet for audit logs, team chat tools like Slack or Discord, a mobile-friendly Telegram message, and a formatted email for executives. Because it deliberately splits the work across several different AI models chosen for specific strengths rather than one general model, the system is described as avoiding the kind of confusion that can come from asking a single model to do everything at once. The README does not state a license for this project.