Your off-market scouting crew

It is a pack of Claude agents: one orchestrator plus scouts tuned per territory and asset class (industrial, multifamily, retail and office). The scouts screen owners against your deal box, score them on a signal stack, resolve entities to real people, and merge one ranked, deduped deal table with a human review gate.

Install the pack in Claude, Claude Code, or ChatGPT using the included README. The orchestrator fans out the per-lane scouts in parallel, each writes into a shared deal-table contract, and it returns one merged, re-ranked table you review before first contact.

It helps but is not required. The flagship pack is the single-agent method: the signal stack, the scoring model, and the honest ceilings. This crew runs that same method in parallel across a whole footprint. If you want the reasoning behind the scores, read the flagship first; if you just want the team running, this pack is self-contained with its own playbook PDF.

Every asserted name carries a documented chain from entity to human, and anything without a clean chain, or resolved at low confidence, goes to a review queue instead of your call list. The scouts read public records and public signals, so accuracy tracks the quality of the underlying county data. Nothing leaves for a letter or a call until a human clears it.

It is free to download and yours to run. The catch is honest: it is a pack you install and operate yourself, not a hosted service. If you would rather we stand it up on your pipeline and run the coverage for you, that is the paid engagement you can book a call about, but the pack itself costs nothing.