DevGuardian – AI Codebase Health Assistant

DevGuardian is an AI-powered Codebase Health & Fix Assistant designed to make software debugging faster, more explainable, and safer. Developers often spend significant time moving between error logs, source files, debugging tools, code changes, and test runners. This disconnected workflow makes it difficult to understand the true root cause of an issue and verify whether a fix is actually safe. DevGuardian brings this workflow into a single controlled pipeline: Issue/Stack Trace → Repository Scan → Relevant Files → Root Cause Analysis → Fix Plan → Patch/Diff → Approval → Regression Tests → Validation → Technical Report A developer can provide a local or controlled Git repository along with an issue description, stack trace, or failing command. DevGuardian scans the repository using deterministic techniques such as file analysis, search, Git information, and Python AST inspection. It identifies the files and code sections most relevant to the reported problem while filtering sensitive files and unnecessary context. The AI then produces a structured root-cause analysis containing the suspected cause, supporting evidence, affected files, confidence, uncertainties, and recommended fix. Before modifying the code, DevGuardian presents a clear fix plan describing the proposed changes, tests, validation steps, and risks. The system generates a reviewable unified diff rather than silently replacing source files. The developer must approve the proposed modification before it is applied. DevGuardian can then generate focused regression tests and edge-case tests and execute only approved validation commands such as Pytest or Ruff. The final dashboard presents the validation results and generates a technical report containing the complete workflow and audit trail. The primary target users are student developers, individual developers, and small development teams who need a simpler and more reliable debugging workflow. What makes DevGuardian unique is its combination of AI reasoning with deterministic developer tools and human approval. AI is used for reasoning, root-cause analysis, planning, and test generation, while repository scanning, Git diffs, test execution, and validation provide controlled and auditable operations. The project focuses on a reliable end-to-end workflow rather than broad language support, demonstrating how AI can turn a bug report into a transparent, test-validated fix.