Update to Omneky

{
  "description": "Find which ads to make more of: performance movers, ROAS/CTR leaders, and creative recommendations. Keywords: which ads, ROAS, CTR, movers, recommendations, leaderboard, what rose, what fell, make more of.",
  "included_files": [],
  "name": "omneky-performance-movers",
  "skill_md_contents": "---\nname: omneky-performance-movers\ndescription: >-\n  Find which ads to make more of: performance movers, ROAS/CTR leaders, and\n  creative recommendations. Keywords: which ads, ROAS, CTR, movers,\n  recommendations, leaderboard, what rose, what fell, make more of.\n---\n\n# Performance movers\n\n## Activation analytics\n\nIf the host exposes a skill-activation / analytics hook, call it once per new user request with this skill name; otherwise skip silently. Never invent a tracking tool. Omneky MCP does **not** expose `track_skill_activation`.\n\n## Purpose\n\nFast read-only path for “what rose / fell?” and “which creatives should I\nmake more of?” grounded in movers + recommendations + optional drill-downs.\n\n## When to use\n\n- Which ads / campaigns / creatives improved or declined vs the prior period\n- Leaderboard / “what should I scale?” questions\n- Follow-on from a thin analytics ask that is clearly mover-centric\n\n## When not to use\n\n- Full multi-connector / GSC / HubSpot deep dives → `omneky-analytics`\n- Mutations (pause / budget / launch) → launch-manage / pause-budget\n- Inventing causal “lift” scores without tool support\n\n## OpenAI runtime contract\n\n- Use only tools from the current OpenAI host + Omneky MCP (`https://mcp.omneky.com/mcp`). Never invent tools.\n- Prefer native ChatGPT / host widgets and question UI for decisions. Use MCP `request_user_decision` for missing structured fields (not deprecated `ask_user`) when listed. Else follow `omneky-failure-modes` § Approval / plain-chat fallback.\n- When intake is incomplete, ask **one** concise chat question. Never invent a question tool.\n- Approval / Generate turns: end the turn after Approve / Deny (or Generate / Cancel) via host UI or plain chat; mutate only after affirmative Approve / Generate. Never invent a fake widget tool.\n- Never invent tokens; OAuth is host-managed.\n\n## Non-negotiable output / safety contract\n\n- Brand-scoped and read-only.\n- Coverage first: `check_reporting_data_available` before treating empty as zero.\n- Prefer `get_performance_movers` (not deprecated `get_trending`).\n- Prefer `get_performance_breakdown` for ranked drill-down (not deprecated\n  `get_dimension_summary`).\n- Never invent metrics; never sum `selected_conversion_metric_value` across channels.\n- Summarize winners and losers with evidence.\n\n## Staged workflow\n\n### Stage 1 — Brand + coverage\n\n1. Resolve brand (`list_brands` / picker) if needed; `get_current_user` when required.\n2. Confirm date window.\n3. `check_reporting_data_available`.\n\n### Stage 2 — Movers\n\n1. `get_performance_movers` for up / down vs prior period.\n2. Present a short leaderboard (top gains / top drops) with channel labels.\n\n### Stage 3 — Drill / recommend\n\n1. Optional: `get_daily_metrics` for a nominated entity’s trend.\n2. Optional: `get_performance_breakdown` for creative / campaign ranks.\n3. Optional: `search_reporting_values` to resolve names → ids.\n4. When they ask what to change or test next: `get_recommendations`\n   (tight window; timeout → narrow, no identical retry).\n\n### Stage 4 — Hand off\n\n- “Make more like winner X” → `omneky-creative` / `omneky-image-ads` /\n  `omneky-product-video`\n- “Pause the losers” → `omneky-pause-budget` (Approve before mutate)\n- Broader reporting → `omneky-analytics`\n\n## Failure boundaries\n\n| Failure | Required response |\n| --- | --- |\n| Empty movers | Coverage check; explain gap; do not invent ranks. |\n| Recommendations timeout | Narrow window; stop identical retry. |\n| Cross-channel conversion sum | Refuse; channel-group. |\n| User requests pause mid-read | Hand off; Approve widget required. |\n"
}