Hapsland

Hapsland catches questionable data and code decisions while your coding agent is still working. It auto-reviews your agent's changes and sends findings back in real time, so the agent can fix a mistake before it builds more code on top of it. It collects semantic context from the diff, follows references to related definitions including those in other files, and reviews each rule against that gathered context. Reviews are batched and stale ones discarded to protect the agent's attention, no prompts or unrelated changes are sent to the classifier, and you can see exactly what goes to the classifier and what comes back.