Write LinkedIn Posts and Comments in Your Voice

Drafts LinkedIn posts, comments, and replies in your voice with approval gates before publishing. It handles LinkedIn's technical quirks (3 URN types, thread flattening, reaction bugs), strips AI tells, and integrates with Publora to publish directly to your account. No coding required.

Once installed, activate any skill by asking Claude about LinkedIn. 1. Provide a topic, draft, or post URL. 2. The skill reads the content, applies 2026 algorithm heuristics and voice rules, then shows you a draft. 3. You approve or edit. 4. On approval, the skill publishes via Publora (or returns copy-paste text if you prefer manual posting). Output: published posts and comments in your voice with drafts reviewed before going live.

Every LinkedIn touchpoint reflects your voice and stays informed by your research, because drafting and approval happen before publication.

Requires Python 3.8+ and requests + python-dotenv packages (installed via pip install -r requirements.txt) to enable Publora auto-posting. Apify token is optional for read-side operations; skills fall back to asking you to paste text if no token is set. Works in Claude Code, Claude Desktop, claude.ai web, Codex CLI, and as a mounted plugin in OpenClaw or custom agents (mount skills/*/SKILL.md files and import lib/ as Python). Publora handles LinkedIn's three URN types and reaction-type mapping automatically.

Publora free tier limits 15 LinkedIn posts/month. Apify optional read-side integration costs $1-5 per 1,000 results. LinkedIn thread flattening means replies to replies must reference the top-level comment URN, not the reply's URN (skills handle this automatically). Some rare post URNs return activity type when the canonical is ugcPost; skills have a fallback.