Mohammad Alsadiq@mohammadalsadiq

PROBLEM & CHALLENGE Human code reviews are one of the biggest bottlenecks in modern software engineering. A single reviewer rarely has the time or cross-domain expertise to deeply audit a Pull Request for security vulnerabilities, database performance regressions, and architectural boundary violations in a single pass. Meanwhile, generic single-prompt AI code wrappers produce noisy, unfocused commentary without actionable fixes, creating tedious back-and-forth cycles for developers. OUR SOLUTION The Multi-Persona Architectural Code Reviewer is an autonomous, agentic Pull Request review committee built with IBM Bob IDE. Instead of passing isolated snippets to a single generic prompt, our platform ingests unified Git diffs and dispatches them concurrently to three specialized AI subagents: 1. SecBot (Security Specialist): Audits for SQL injection vectors, hardcoded secrets/tokens, authentication bypasses, XSS, and command injection. 2. PerfBot (Performance Specialist): Detects N+1 database queries, unindexed full-table scans, blocking I/O inside loops, and O(N^2) memory bloat. 3. ArchBot (Architecture Specialist): Flags modular layer violations (such as UI components directly querying database models), tight coupling, and breaking API contracts. TARGET USERS & HOW THEY INTERACT WITH IT Built for software engineers, tech leads, and DevSecOps teams, users interact with a split-screen web dashboard (Next.js/Tailwind + FastAPI): - Developers paste a raw Git diff, load pre-built compliant sample patches, or import a live public GitHub PR/commit URL (with automated PII header scrubbing). - Clicking "Run Multi-Persona Review" triggers all three subagents in parallel, returning a 0-100 Code Health Score, merge recommendation, and categorized issue cards. - Developers can click "Apply AI Suggestion" on any finding to preview and copy a before-and-after unified diff refactor patch, watch the Health Score recover in real time, chat 1-on-1 with each specialist robot, or query all three agents simultaneously in the Review Council Console. WHAT MAKES IT CREATIVE, UNIQUE & EFFICIENT By combining Python asyncio.gather parallel subagent orchestration, cross-agent finding deduplication, one-click executable refactor patches, and an interactive multi-agent Council Console, our solution cuts review latency by nearly 3x compared to sequential analysis while turning passive critique into instant code repairs.

💻 Online Hackathon | ⏳ Build: 48 hours | 📅 September 25–27, 2026 The official IBM Bob 2.0 hackathon. Build what's next in AI-assisted development, alongside the developers and experts shaping it. ⭐ Get help from experts throughout the event 🤝 Join a team or build solo 💰 Prize pool: $12,000 🧑🏻‍💻 Register before the Kick-Off Stream to start building with Bob