gtm-pipeline:outreach

Use the LinkedIn Auto Connect PhantomBuster agent with a generated Python script that launches the agent with a one-off argument override. The script processes one profile per run and tracks progress in a state file, so you can loop it every 20-30 minutes safely.

Use the Sales Navigator Message Sender PhantomBuster agent with a CSV containing linkedin_profile_url, linkedinMessage, and inmailSubject columns. Generate pb_message.py from the template, test it once, then run it on a timed loop until all profiles are processed.

LinkedIn enforces a hard limit of about 100 connection requests per week, so pace Auto Connect at one profile per run and roughly 14 per day. Message sending can run 1-3 profiles per run, with InMail credits applying for third-degree or non-connections.

The error 'numberOfAddsPerLaunch => is the wrong type' means the argument JSON contains string values where booleans or integers are expected. Fix it by ensuring json.dumps() is applied to the argument dict in the launch request body.

You need a CSV with a linkedin_profile_url column for every contact, plus linkedinMessage and inmailSubject columns if you plan to send messages or InMails. Typically this comes from a people-enrichment step that outputs contacts_enriched.csv.