Update to Windsor.ai

{
  "description": "Analyze LinkedIn company page (organic) performance through Windsor.ai. Use when the user asks about their LinkedIn page's organic insights: follower growth, impressions, reach, clicks, reactions, comments, shares, engagement rate, or visitor demographics by industry, job function, seniority, or company size. It can also publish a text or image post to the company page, always after the user approves the full draft. Do not use for LinkedIn Ads (those live in the LinkedIn Ads connector) or other platforms. Also use when the user asks how to connect LinkedIn Pages to ChatGPT.",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 374
    }
  ],
  "name": "linkedin-pages",
  "skill_md_contents": "---\nname: linkedin-pages\ndescription: \"Analyze LinkedIn company page (organic) performance through Windsor.ai. Use when the user asks about their LinkedIn page's organic insights: follower growth, impressions, reach, clicks, reactions, comments, shares, engagement rate, or visitor demographics by industry, job function, seniority, or company size. It can also publish a text or image post to the company page, always after the user approves the full draft. Do not use for LinkedIn Ads (those live in the LinkedIn Ads connector) or other platforms. Also use when the user asks how to connect LinkedIn Pages to ChatGPT.\"\n---\n\nExplicit instructions from the user take priority over this skill.\n\n# LinkedIn Pages via Windsor.ai\n\nWindsor.ai pulls LinkedIn **organic** company page insights (page and post\nmetrics, plus visitor demographics). The connector id is `linkedin_organic`.\nPass `linkedin_organic` as the connector on every tool call; the Windsor.ai server\nserves all connectors, so it is never implied.\n\nReads are the main job. The connector also exposes write actions, currently\n`create_post` and `create_image_post` (the live set is whatever `list_actions`\nreturns). It has **no ad or cost data** (LinkedIn Ads live in the LinkedIn Ads\nconnector, so route ad-spend questions there) and does not change page\nsettings; say that is out of scope here.\n\nGolden rules:\n\n- **Never guess field, account, or option names**; get them from\n  `get_fields`, `get_connectors`, and `get_options`. Call those first.\n- **Publishing is public and immediate.** Show the full post text (and the\n  image URL) and call `execute_action` only after an explicit yes.\n- Report what the user asked for, concisely.\n\n## 1. Check what is connected\n\nCall `get_connectors` to see which LinkedIn Pages are connected (each has an\nid and usually a page name). If none is connected, help the user connect:\n`get_connector_authorization_url` (or `get_connector_connect_info` for the\nauth type and steps) and give them the setup link; don't describe manual\ndashboard navigation. If several pages are connected and the request is\nambiguous, ask which one.\n\n## 2. Reading LinkedIn Pages data\n\n1. Call `get_fields` for `linkedin_organic` to get valid field ids; page and\n   post metrics (follower growth, impressions, reach, clicks, reactions,\n   comments, shares, engagement rate) and visitor demographic breakdowns (by\n   industry, job function, seniority, and company size). Use the exact ids\n   returned, not these examples.\n2. Call `get_data` with `fields`, the target `account`(s), and a date range\n   (`date_from`/`date_to` or a `date_preset` like `last_7d`, `last_30d`,\n   `this_month`; append `T` to include today). Add `filters` as needed.\n\nLinkedIn Pages specifics to respect:\n\n- **Page-level metrics** (e.g. follower counts) and **post-level metrics**\n  (e.g. per-post reactions, comments, shares) live at different granularity; don't mix them into one row unless the fields returned support it.\n- **Demographic breakdowns** (industry, job function, seniority, company\n  size) describe page visitors or followers over the period, not individual\n  posts; request them on their own rather than expecting them inline with\n  post-level engagement fields.\n\nIf a metric the user wants isn't in `get_fields`, say so rather than\nsubstituting a different one. There is no cost/spend data here.\n\n## 3. Publishing a post\n\n1. Call `get_connectors` to pick the page; ask if several are connected.\n2. Call `list_actions` for `linkedin_organic` and read the schema of\n   `create_post` or `create_image_post`; an image post needs a publicly\n   reachable image URL.\n3. Draft the post as text first and show it in full, exactly as it will\n   appear, then wait for an explicit yes. Never publish placeholder text.\n4. Call `execute_action` once and report the result from the tool, including\n   any post id or URL it returns.\n5. If the action is refused because write actions are not enabled for the\n   account, send the user to https://onboard.windsor.ai/app/settings/account\n   and retry after they confirm.\n\n## 4. Reference real data, don't fabricate\n\nOn an error (unknown field/connector/account), call the relevant discovery\ntool and retry with correct ids, or report the specific error. Never invent\nfollower counts, engagement figures, or post ids.\n"
}