A crawler that keeps facts separate from claims

Crawler agent prototype fetches approved links and extracts timestamped facts.

The crawler agent prototype has one narrow job: fetch approved public links and pull out timestamped facts. It doesn't decide what those facts mean for someone's profile. That separation is deliberate. Extracted evidence gets normalized before it ever reaches the professional memory that feeds a profile, and the step of turning evidence into a user-facing claim happens elsewhere, under review. Keeping fact extraction and claim-making apart makes it easier to trust either piece on its own, and easier to spot where a claim might be running ahead of what the evidence actually shows.