company-research

name	company-research
description	Company research using Exa search. Finds company info, competitors, news, financials, LinkedIn profiles, builds company lists. Use when researching companies, doing competitor analysis, market research, or building company lists. Triggers on "research company", "competitors of", "companies in", "market map".
context	fork

MCP server note (this workspace): Exa tools come from the exa-search MCP server — mcp__exa-search__web_search_advanced_exa and mcp__exa-search__deep_search_exa. Load them via ToolSearch if not already available.

Company Research

Tool Restriction (Critical)

This skill calls web_search_advanced_exa. For deep multi-angle synthesis (query expansion + cited answer), hand off to the separate deep_search_exa tool — that is a different tool, not a mode of this one. Note that type: "deep" is a param of web_search_advanced_exa (a search depth), distinct from the standalone deep_search_exa tool. Don't reach for unrelated Exa tools from here.

Token Isolation (Critical)

Never run Exa searches in the main agent context for bulk work. Always spawn a Task agent:

• Task agent runs the Exa search internally
• Task agent processes results using LLM intelligence
• Task agent returns only distilled output (compact JSON or brief markdown)
• Main context stays clean regardless of search volume

For a single small lookup (one query, < 10 results), running inline is fine.

Dynamic Tuning

No hardcoded numResults. Tune to user intent:

• "a few" → 10–20
• "comprehensive" → 50–100
• User specifies a number → match it
• Ambiguous → ask: "How many companies would you like?"

Query Variation

Exa returns different results for different phrasings. For coverage:

• Generate 2–3 query variations
• Run them in parallel
• Merge and deduplicate

Categories

Pick category based on what you need:

• company → homepages, rich metadata (headcount, location, funding, revenue)
• news → press coverage, announcements
• people → public LinkedIn profiles
• omit / type: "auto" → general web results, broader context

Start with category: "company" for discovery, then drill down with other categories.

Category-Specific Filter Restrictions

When using category: "company", these parameters cause 400 errors:

• includeDomains / excludeDomains
• startPublishedDate / endPublishedDate
• startCrawlDate / endCrawlDate

When searching without a category (or with news), domain and date filters work fine.

type is orthogonal to category: the 400 restrictions key off the category, not the search type.

Universal: includeText / excludeText only support single-item arrays. Multi-item arrays cause 400 errors across all categories.

LinkedIn

Public LinkedIn via Exa: category: "people", no other filters. Auth-gated LinkedIn → fall back to Claude in Chrome.

Examples

Discovery

web_search_advanced_exa
{
  "query": "AI infrastructure startups San Francisco",
  "category": "company",
  "numResults": 20,
  "type": "auto"
}

Deep dive on one company

web_search_advanced_exa
{
  "query": "Anthropic funding rounds valuation 2024",
  "type": "deep",
  "numResults": 10,
  "includeDomains": ["techcrunch.com", "crunchbase.com", "bloomberg.com"]
}

News coverage

web_search_advanced_exa
{
  "query": "Anthropic AI safety",
  "category": "news",
  "numResults": 15,
  "startPublishedDate": "2024-01-01"
}

LinkedIn profiles

web_search_advanced_exa
{
  "query": "VP Engineering AI infrastructure",
  "category": "people",
  "numResults": 20
}

Output Format

Return:

1. Results — structured list, one company per row
2. Sources — URLs with a one-line relevance note
3. Notes — uncertainty or conflicts

---

Hand-off: to persist distilled findings into raise state, route through the round CLI — round investors update / round outreach draft — as proposals only. Never write round.db directly, never pass --yes, never reference local.db. This skill stays read-only research; persistence is a hand-off.