Update to Klue

{
  "description": "Use for Klue questions that need fresh evidence beyond a sufficient curated answer: competitive research, source-backed claims or quotes, account and opportunity investigation, recent changes, win/loss analysis, counts, trends, or comprehensive comparisons.",
  "included_files": [],
  "name": "research",
  "skill_md_contents": "---\nname: research\ndescription: Use for Klue questions that need fresh evidence beyond a sufficient curated answer: competitive research, source-backed claims or quotes, account and opportunity investigation, recent changes, win/loss analysis, counts, trends, or comprehensive comparisons.\n---\n\n# Klue Research\n\nGather the minimum sufficient evidence for the user's question, then stop.\nUse [citation-instructions](../citation-instructions/SKILL.md) when turning\nthat evidence into an answer.\n\n## Select the current platform workflow\n\nInspect the tools available in the current session and their descriptions. A\nclient exposes one platform surface; never infer its platform from the user's\nwording and never combine workflows.\n\n| Tool signals | Workflow |\n| --- | --- |\n| `Search Klue` / `klue-search-content` with Smart Answers or 2.0 deal-support tools | **Compete 2.0** |\n| `Search Klue` / `klue-search-content` limited to win/loss, without 2.0 Smart Answer/deal-support tools | **Klue 1.0** |\n| Neither recognizable surface | Use only tools whose descriptions clearly fit the request; explain the limitation if none do. |\n\nSearch content is the default research path on both surfaces. Its result-source\nlabels, including Stack Cards, are not platform signals. The categories and\nfilters in its live description are an access boundary.\n\n### Compete 2.0\n\n- Use `Search Klue` / `klue-search-content` as the default research path. Use\n  its live description for supported sources, filters, and parameters.\n- Check Smart Answers first when available. They can answer a stable,\n  well-matched question directly, but are not sufficient alone for a\n  time-sensitive, account-specific, source-specific, conflicting, or partially\n  covered request.\n- For a named account, opportunity, buyer, prospect, renewal, or recent deal:\n  resolve the opportunity when needed, then retrieve the available buyer\n  requirements, competitors, and recent deal evidence. Supplement with unified\n  search when it materially answers the requested angle.\n\n### Klue 1.0\n\n- Use the exposed `Search Klue` / `klue-search-content` capability for\n  competitive research. Klue 1.0 exposes win/loss content only; use only the\n  categories, filters, and fields documented by that live tool description.\n- A rejected category is an access boundary, not a query to retry. Do not\n  borrow 2.0-only fields or parameters.\n\n## Plan the evidence\n\nBefore searching, silently identify the required entities, requested source\ntypes, date window, deal context, and deliverable. Preserve the user's angle;\ndo not expand a focused question into adjacent topics merely to collect more\nmaterial.\n\n- **Targeted research:** one or two named angles, a specific claim, a focused\n  source request, or a single account/deal.\n- **Broad research:** an open comparison, multiple competitors, several\n  independent angles, a pattern, roster, ranking, or explicitly comprehensive\n  request.\n- **Follow-up research:** only to close a named evidence gap, verify recency,\n  resolve a conflict, or retrieve a requested quote/detail.\n\nRun independent searches together when the connector supports it. Keep\ndependent steps ordered: resolve an opportunity before requesting its details;\nconfirm a filter value before relying on it when the tools provide metadata\nenumeration.\n\n## Retrieval rules\n\n- Query each source type the user explicitly requests. If the request asks for\n  buyer voice, search for buyer evidence rather than only enablement material\n  or a competitor's own claims.\n- Use primary source material to validate decisive claims when available.\n  Summaries and synthesized insights can orient research but do not turn into\n  verbatim buyer evidence.\n- For a \"latest\" claim, compare explicit dates within the requested scope and\n  check for a newer candidate when the capability supports it. Otherwise say\n  \"newest found,\" not \"latest.\"\n- For a trend or change claim, obtain evidence from both the earlier and later\n  periods. One time slice does not prove a trend.\n- Do not guess tool parameters, filter names, category values, IDs, or source\n  coverage. Current tool descriptions and returned schema are authoritative.\n- If a valid filtered search returns no results, widen one level at a time:\n  remove the narrowest subcategory filter, then the category filter, then run\n  the query unfiltered. Preserve the user's query and stop once evidence is\n  sufficient. Do not widen a rejected category, invalid parameter, or\n  permission-bound query.\n- Use public web research only when the user asks for public sources or a\n  current public fact is unavailable in Klue. It never substitutes for\n  proprietary deal, call, message, or win/loss evidence.\n\n## Stop conditions\n\nStop when the core question has enough citable evidence and any material caveat\nis known. More search is not automatically better.\n\nIf a required part remains unsupported after a bounded gap-fill, answer the\nsupported parts and name the missing evidence. Do not manufacture a complete\nanswer from near matches.\n"
}