find-sellers
Get homes from a data source or spreadsheet, score them, and show Steve the best off-market seller leads. Use when Steve asks for leads, new homes, a list, the report, owner emails, or to load a file.
SKILL.md.claude/skills/find-sellers/SKILL.md
GitHubFind likely sellers
- Run
uv run python -m sellers sourcesto see which sources are ready. - Pick the source with Steve:
- demo: fake homes for practice. Free.
- rentcast: real owners. Each city costs 1 request, and the free plan has
50 requests a month. Before you run it, tell Steve how many requests this
pull uses (one per city in
settings.toml). - a spreadsheet (Realist, PropStream...): ask Steve where the file is,
then run
uv run python -m sellers import "<path>". If the address column is not found, read the column names it printed. Add the right one toALIASESinsellers/sources/csv_file.py, then run it again.
- Pull:
uv run python -m sellers pull <source>. Add--city "Dana Point"for one city. The tool scores the homes for you. - Show the top homes:
uv run python -m sellers top --n 10. Explain in plain words why the first 2-3 homes scored high. - Open the full list:
uv run python -m sellers report. Tell Steve he can search, filter, and click the column names to sort.
Next steps to offer, one at a time:
- No emails yet? Offer to find them (see below).
- Has emails? Offer
/write-emails. - Wants a different order? Change the points in
[scoring]insettings.toml, then runuv run python -m sellers score. - A home is not a fit? Run
uv run python -m sellers mark <id> skip --note "why".
Finding owner emails (Tracerfy)
- Practise first, for free:
uv run python -m sellers find-emails --test --n 3. The results are fake and use example.com. - Live lookups cost about 2-4 cents per owner found, and nothing if no one is
found. Before each live run, ask Steve and tell him the most it can cost
(
--ntimes $0.04). Then runuv run python -m sellers find-emails --n <number>.
Never share owner details outside this computer, except in Steve's own Gmail drafts.