LinkedIn (1)

**Quick judgment**: This is a powerful, general-purpose LinkedIn integration for agents that need real read/write access to LinkedIn (posts, media, org data, and advertising) plus public Ad/Job Library search. It is suitable for production use if you are comfortable depending on the Maton gateway and you enforce strong safeguards around posting and ad spend. **Recommended scenarios** - Building an AI social media manager that can draft and publish LinkedIn content, including images and videos, for users or organizations. - Marketing/ads copilots that manage LinkedIn ad accounts, campaign groups, and campaigns, including budget and status changes. - Analytics, research, or recruiting assistants that query the Ad Library and Job Library, and retrieve organization info and posts. **Key risks & considerations** - **Account & content safety**: The skill can publish posts and manage ads; agents must not post or modify campaigns without clear, explicit user intent and confirmation (e.g., dry-run/preview flows, user approval steps). - **Financial risk**: Ad account and campaign management implies potential spending; agents should surface budgets, currencies, and statuses clearly and require user confirmation before creating or activating campaigns. - **Privacy & compliance**: All access is through Maton (gateway.maton.ai / ctrl.maton.ai) using a MATON_API_KEY and LinkedIn OAuth tokens. You must treat keys and tokens as sensitive, respect LinkedIn’s terms of service, and avoid scraping-like behavior or excessive calls. - **Operational dependency**: Uptime and behavior depend on both LinkedIn’s API and Maton’s gateway; agents should handle HTTP errors, expired connections, and rate limits gracefully. - **Complexity for non-experts**: While well-documented, it largely exposes the raw LinkedIn REST surface; agents should encapsulate common patterns (e.g., “create text post”, “search ads by keyword”) rather than blindly constructing arbitrary payloads.