ShipGuard AI - software release readiness agent

PROBLEM: Software projects can contain hidden problems even when the application appears to work. Inconsistent API documentation, incorrect environment configuration, incompatible runtime versions, failing tests, incomplete deployment documentation, and other overlooked issues can cause failures after deployment. For small development teams, and students in particular, manually reviewing all of these areas before release can be time-consuming and easy to overlook. THE SOLUTION: ShipGuard AI is an AI-powered release-readiness and software audit agent that reviews a software project against its defined release requirements. The agent examines project files such as source code, configuration files, package information, tests, Docker configuration, documentation, and release requirements. It maps requirements to evidence in the project and identifies problems as blockers, warnings, or passing requirements. Instead of only reporting problems, ShipGuard AI can generate a structured fix plan showing the affected file, the issue, the proposed change, and the requirement being addressed. This allows developers to review proposed changes before applying them. HOW USERS INTERACT WITH IT: A developer provides ShipGuard AI with a software project and its release requirements. The system analyzes the project and produces an audit report showing: * Requirements that pass * Release blockers * Warnings * Evidence supporting each finding * Proposed fixes * Verification results The developer remains involved in the process by reviewing and approving proposed changes before they are applied. WHAT MAKES IT UNIQUE: ShipGuard AI goes beyond a traditional code checker. It connects requirements with actual project evidence and evaluates the project as a release package rather than looking at source code alone. For example, during our FoodNest demonstration project, ShipGuard identified issues involving payment behavior, environment-variable configuration, Node.js version consistency, API documentation, and database migration documentation. It then generated targeted fixes for the identified issues. This requirement-to-evidence approach makes the system useful for catching problems that may exist between different parts of a project, such as when the README documents one API endpoint while the server implements another. EFFECTIVENESS AND EFFICIENCY: ShipGuard AI reduces the amount of manual checking required before release by systematically reviewing multiple project components and organizing its findings into an actionable release report. It also separates blockers from warnings, helping developers understand which issues require attention before release and which issues represent remaining limitations. In our prototype validation, the FoodNest project's automated tests passed after the relevant fixes were applied. The project also demonstrates transparent handling of limitations: because FoodNest does not currently contain a database, ShipGuard documents the database migration requirement as a warning rather than falsely claiming that a production database migration has been implemented. Overall, ShipGuard AI aims to make software releases more systematic, transparent, and reliable by turning release requirements into an actionable engineering audit.