Sentinel is an agentic security analyst for developers and bug-bounty hunters. Paste code, drop in files, or just paste a GitHub repo URL, and Sentinel runs a real, multi-step review inside a sandboxed agent. You get back a security grade and triaged findings (severity, CWE, the exact file and line, impact, and the fix), a shareable HTML report, and machine-readable JSON. Then one click asks the same agent for the hardened code. The agent backend is a single HarnessRouter harness running Qwen Code CLI on qwen3.8-max, a model that will reason through an exploit path in code you own or a target you're authorized to test, instead of refusing. It's the difference between "I can't help with that" and a prioritized bug list you can act on this afternoon.