← ClaudeAtlas

network-mapperlisted

Find warm-intro paths into a target company from the user's own LinkedIn connections export. First-degree from the CSV; second-degree via enrichment + human-in-the-loop. No scraping. Scheduled weekly or on demand.
HimadriTrying/ai-job-search-agent · ★ 1 · AI & Automation · score 68
Install: claude install-skill HimadriTrying/ai-job-search-agent
# network-mapper — the gap nobody else builds ## Legal stance LinkedIn blocks automation and litigates scraping. This skill uses **only the user's own data**: LinkedIn Settings → Data Export → Connections → CSV in `data/connections/`. No scraping, no ToS violation. ## First-degree (fully automatable today) 1. Parse the connections CSV. 2. Given a target company, find who in the network works (or worked) there. 3. Rank by relationship strength signals available in the export (connection date, shared context if present). ## Second-degree (the genuinely hard part — optimise the human step, don't fake it) The export is first-degree only. Options, in order of preference: - **Enrich** first-degree connections' current employers (Harmonic `enrich_person`, or Apollo/Hunter via API) and reason about who is *likely* to know someone relevant. - **Ask** — draft a message for the user to send a first-degree contact requesting an intro. Often the export can't prove the second-degree link, so make the human ask efficient rather than eliminating it. ## Output For each target role: warm-intro paths found, ranked, each with a suggested next action. Hands off to `outreach-drafter`. Scheduled weekly (network changes slowly) or on demand.