Update to Revelata deepKPI

{
  "description": "Answer \"what does this company do?\" and segment / geography questions using deepKPI: get_company_summary (narrative summary) and get_company_segments (segment & operating structure). For discovery — \"which companies do X\", \"who operates in Y\" — use company_summary_search across summaries, then pull full summaries for names that need detail. NOT for numeric KPI time series — use the kpi skill for that. Triggers: what does X do, describe the business, segments, geographies, reportable segments, companies that make, firms in industry, thematic lists.\n",
  "included_files": [],
  "name": "company-summary",
  "skill_md_contents": "---\nname: company-summary\ndescription: >\n  Answer \"what does this company do?\" and segment / geography questions using deepKPI:\n  get_company_summary (narrative summary) and get_company_segments (segment & operating\n  structure). For discovery — \"which companies do X\", \"who operates in Y\" — use\n  company_summary_search across summaries, then pull full summaries for names\n  that need detail. NOT for numeric KPI time series — use the kpi skill for that.\n  Triggers: what does X do, describe the business, segments, geographies,\n  reportable segments, companies that make, firms in industry, thematic lists.\n---\n\n# company-summary\n\n**Purpose:** Ground answers in deepKPI's **10-K-derived company summary** and **segment\nbreakdown** endpoints. Use **summary search** when the user is looking for **names**\nthat match a **theme** or **activity**.\n\nFor **historical metrics**, provenance tables, and Excel-style pulls, use the `kpi`\nskill instead.\n\n## Data source\n\nSee `../_common/data-source-rules.md` for the runtime routing table and the canonical\ninventory of deepKPI operations. This workflow uses `query_company_id`,\n`get_company_summary`, `get_company_segments`, and `company_summary_search`.\n\n`query_company_id` resolves **name / brand → `company_id`** (SEC CIK for US issuers).\nUse it whenever the user names a company and you do not already have `company_id`.\n\n## Hard stop if deepKPI access fails\n\nSee `../_common/connection-failure.md`. STOP and ask before falling back.\n\n## Credits (budget consciously)\n\n| Call | Cost |\n|------|------|\n| **`get_company_summary`** | **3 credits** on success (one company, latest 10-K narrative). |\n| **`get_company_segments`** | **3 credits** on success (one company, structured segments / geos from latest 10-K). |\n| **`company_summary_search`** | **1 credit per company returned** (`top_k_companies`, max **15**). Empty results: **0**. |\n\n## Workflow A — Named company (default)\n\nUse this when the user is asking about **one or more specific issuers** by name or\nticker.\n\n### What does the company do? (business, model, strategy)\n\n1. **`query_company_id`** if you need **`company_id`**.\n2. **`get_company_summary`** — return the **narrative** from the response; that is the\n   canonical answer to \"what do they do?\", \"describe the business\", \"how do they\n   make money at a high level?\", etc.\n\nDo **not** substitute web summaries or training-data descriptions when this endpoint\nis available.\n\n### Segments, geographies, how they break out the business\n\n1. **`query_company_id`** if needed.\n2. **`get_company_segments`** — use for **segment structure**, **operating geographies**,\n   **reportable segments**, and similar **organizational** questions (what lines /\n   regions they disclose), as returned by the API.\n\nFor **segment revenue / margins over time**, follow the `kpi` skill (`list_kpis` /\n`search_kpis`) after you have the segment names you care about.\n\n### Both narrative and structure\n\nCall **`get_company_summary`** and **`get_company_segments`** when the question needs\n**prose plus** how they **slice** the business (same `company_id`).\n\n## Workflow B — Thematic discovery (no single target company)\n\nUse **`company_summary_search`** when the user wants a **set of companies** defined by\n**what they do** or **where / how they operate**, e.g.:\n\n- \"Companies that **make** …\", \"**operate in** …\", \"**exposed to** …\"\n- \"Who **reports** …\", \"**US-listed** firms in …\" (as natural language — tune the `query`)\n\nSteps:\n\n1. **`company_summary_search`** with a clear natural-language **`query`** and\n   **`top_k_companies`** (often **10**; max **15**). **1 credit per row returned.**\n2. Present the ranked list (names / `company_id` / ticker as returned).\n3. If the user (or the task) needs **full narrative** for one or more hits — not just\n   the fact they matched — call **`get_company_summary`** on those `company_id`s\n   (**3 credits** each). Use **`get_company_segments`** on selected names if they ask\n   how those companies **break out** operations.\n\nIf results are weak, **rephrase `query`** before giving up.\n\n## Out of scope — benchmarking vs this skill\n\n**Not covered here:** \"**Companies most similar to [Target]**\", formal **peer sets**\nfor valuation comps, or **benchmark** workflows keyed off a **specific** company as\nthe anchor. Use the `benchmark` skill for that. This skill's\n**`company_summary_search`** is for **theme / activity** discovery (\"who does X?\"),\nnot a drop-in replacement for target-based similarity benchmarking.\n\n## Opening line\n\nPick one tone that matches the ask, e.g.:\n\n- Named company: **\"I'll pull their summary and segment breakdown if needed from deepKPI.\"**\n- Thematic list: **\"I'll search deepKPI's company summaries for matches, then pull full summaries where we need detail.\"**\n\n## Common failure modes\n\n- **Answering \"what do they do?\" from memory** instead of **`get_company_summary`**.\n- **Guessing segment names** instead of **`get_company_segments`** (or KPI search) when\n  the user asked how the business is **broken out**.\n- **Using `company_summary_search`** when the user named **one company** and only need\n  summary/segments — use **Workflow A**.\n- **Expecting full segment financial history** from **`get_company_segments` alone** —\n  use the `kpi` skill for time series.\n"
}