conference-scheduler

Run the find command with an industry name or id, and the skill searches the web for conferences, trade shows, and association annual meetings for that niche. Each result is fetched for dates, location, cost, and exhibitor profile, then scored and written to the tracker.

Each event is scored 1-5 on industry relevance (40% weight), owner likelihood (40%), and travel friction (20%, inverted). The composite formula is (Industry x 0.40) + (Owner x 0.40) + ((6 - Friction) x 0.20), with 4.0+ meaning go and below 3.0 meaning skip.

No API key is required. Web search is the only data source; the skill uses WebSearch and WebFetch to discover events and read conference sites for agendas, exhibitor lists, and speaker bios.

Data lives in a shared state directory resolved by lib/paths.py, defaulting to ~/Library/Application Support/nsp-autopilot/ and overridable via the NSP_SEARCH_DIR environment variable. It writes conference-tracker.json as the source of truth and regenerates Conference-Calendar.md as a readable view.

Use the resolve command with a date or month. The skill compares composite scores, weighs industry priority first and owner likelihood second, asks whether each event recurs annually, and records the losing event's decision so the conflict is not re-evaluated later.

No. It is a companion tool, not a stage of run-cycle: it produces no sequences, contacts, or copy. It handles the in-person side of sourcing and shares only the niche slug convention with the industry tracker.