Update to RankZero

{
  "description": "Use RankZero for AI search visibility analysis, brand reporting, competitor research, tracked prompts, citations, and connected GA4/GSC analysis. Use when the user asks about their RankZero data, AI SEO performance, brand visibility in AI answers, competitors, prompts, sources, or related traffic data.",
  "included_files": [],
  "name": "rankzero-workflows",
  "skill_md_contents": "---\nname: rankzero-workflows\ndescription: Use RankZero for AI search visibility analysis, brand reporting, competitor research, tracked prompts, citations, and connected GA4/GSC analysis. Use when the user asks about their RankZero data, AI SEO performance, brand visibility in AI answers, competitors, prompts, sources, or related traffic data.\n---\n\n# RankZero workflows\n\nRankZero is an AI-search visibility platform. Prefer its live MCP data over assumptions when answering questions about a connected brand.\n\n## Brand context\n\n- Start with `list_brands` when the user has not identified a brand or when the brand is ambiguous.\n- Pass `brand` on every call. It accepts a name, domain, UUID, or number from the last brand list.\n- `set_active_brand` sets a default for calls that omit `brand`. That default is best-effort session memory, so prefer passing `brand`.\n- If the user names a domain, resolve it through `list_brands` or pass the domain to a supported brand argument rather than guessing an ID.\n\n## Core analysis\n\nFor a high-level health check, use:\n1. `get_brand_overview`\n2. `get_brand_competitors`\n3. `get_ai_mentions`\n\nOverview and mentions accept `period`: `24h`, `7d`, `30d`, or `90d` (default `7d`). These are rolling windows, not arbitrary historical ranges. State the window and distinguish measurements from interpretation; do not present overlapping windows as consecutive periods.\n\n## Content and AI-search opportunities\n\n- Use `list_tracked_prompts` to list active tracked prompts, 20 per `page`, optionally filtered by `search`.\n- Use `search_knowledge_base` with `query` to search RankZero's documentation, guides, glossary, and blog (top five matches).\n- Use `get_citation_sources` for cited domains by usage. It accepts the same periods (default `30d`), optional `country` and `provider`, and `limit` (default 50).\n- Do not claim that a recommendation will guarantee rankings or AI citations.\n\n## Traditional SEO + AI SEO\n\nWhen the brand has integrations connected:\n- Use `get_gsc_performance` for clicks, impressions, CTR, and position. `dimensions` accepts `date`, `query`, `page`, `country`, `device`, or `searchAppearance` (default `[\"date\"]`). Dates use `YYYY-MM-DD`, defaulting to the last 28 days through today.\n- Use `get_ga4_report` for traffic and behavioral data with chosen `metrics` and `dimensions`. Dates accept calendar or relative values (default `28daysAgo` to `today`).\n- Compare GSC/GA4 metrics with AI visibility only when the date ranges and dimensions are comparable.\n\nIf a Google integration is unavailable, report that limitation instead of inventing the data.\n\n## Reporting style\n\nWhen asked for a report:\n- Lead with the measured metrics.\n- Include the time window.\n- Highlight meaningful changes, competitors, mentions, and sources when the data supports them.\n- Separate observations from recommendations.\n- Keep recommendations concrete: pages, prompts, sources, or technical/content areas to investigate.\n\n## Safety and accuracy\n\nRankZero data is account-scoped. Never imply access to brands the connected account cannot access. Do not expose authentication tokens or ask the user to paste credentials.\n"
}