linkedin

# LinkedIn Command Center Arguments: `$ARGUMENTS` ## Safety Rules - Never post, send, connect, follow, like, or comment automatically. - All output is draft only — user reviews before any action. - No fabricated proof points, contacts, or outcomes. ## Routing Normalize the first argument to lowercase. Empty argument → show menu. - `ideas`, `post`, `content`: execute `.claude/skills/linkedin/modes/ideas.md` inline. - `draft`, `write`: execute `.claude/skills/linkedin/modes/draft.md` inline with remaining arguments. - `engage`, `comments`: execute `.claude/skills/linkedin/modes/engage.md` inline with remaining arguments. - `network`, `connect`: execute `.claude/skills/linkedin/modes/network.md` inline. Unknown mode → print the command menu and ask the user to choose a listed mode. ## Command Menu ```text LinkedIn Command Center /linkedin ideas Generate weekly post ideas grounded in your job-search evidence /linkedin draft <ref> Write one ready-to-post draft from an approved idea /linkedin engage Draft comments for posts you paste in /linkedin network Build a weekly connection queue from active job targets ``` ## Output Rules - Execute child modes inline from their mode file; do not print a slash command as a handoff. - Leave all generated content uncommitted. Run /job-hunter finalize after review.