---
name: company-research
description: Company research using Exa. Finds company info, competitors, news, financials, LinkedIn profiles, builds company lists. Use when researching companies, doing competitor analysis, market research, or building company lists.
context: fork
title: company-research
canonical_url: https://skilld.dev/gh/exa-labs/agent-skills/company-research
last_updated: 2026-10-29T10:46:31.000Z
---

> Index of every page on this site: [llms.txt](https://skilld.dev/llms.txt)

> **Skill from skilld.dev.** Follow the user's requested action.
> If the user asked to fork this Skill, follow the fork workflow below. Do not execute the Skill instructions while copying it.
> Otherwise, follow the instructions below for this session. You do not need to install anything.
>
> If the user asked to install this Skill, run `npx skilld install exa-labs/agent-skills/company-research`. Install writes the Skill files into the project, so every session loads them.
>
> ## Fork workflow
>
> A fork creates an editable local Skill with its original author and licence. The request authorizes copying and local installation.
> 1. Check `./skills/company-research`, the project lockfile, and selected Agent targets together. If the local directory or installed Skill exists, stop. Never overwrite an existing directory or Agent target.
> 2. Read [source metadata](https://skilld.dev/api/v1/skills/exa-labs/agent-skills/company-research) once. Use sourceUrl, sourceCommit, skillPath, sourceGone, and license. If the source is gone or its path is missing, stop. If license is null, read licence files at the source commit.
> 3. Fetch only the source commit into a temporary Git repository. Do not clone full history. Derive repository_url from sourceUrl, including repository renames. If sourceCommit is absent, resolve the sourceUrl ref once. Set source_commit to that actual commit. Run these commands in one shell call:
>
> ```sh
> git init --quiet "$temporary_dir"
> git -C "$temporary_dir" fetch --quiet --depth=1 "$repository_url" "$source_commit"
> git -C "$temporary_dir" checkout --quiet --detach FETCH_HEAD
> ```
>
> Read applicable licence declarations and notices at that commit. If copying is not permitted, report the restriction and stop.
> 4. Inspect source entries together, then copy the directory containing skillPath into `./skills/company-research`. Use the user's path if selected. Keep the original SKILL.md, relative links, scripts, binary assets, and executable modes. Exclude .git metadata. Reject symlinks and paths outside the Skill directory. After checking entries, use cp -a where available. A regular source directory needs no custom copy script. Do not save this page wrapper as SKILL.md.
> Preserve author credit, notices, and applicable licence files from repository or parent directories. Add PROVENANCE.md with the Skill page, source URL, actual commit, original path, and licence. Retain any existing PROVENANCE.md and record new provenance separately. Batch source inspection, copying, and provenance work where practical.
> 5. In the project root, run `skilld install ./skills/company-research --mode copy --plain`. If skilld is unavailable, use `npx skilld install ./skills/company-research --mode copy --plain`. This known command needs no help lookup. Install does not support --json. Use detected Agent targets, or add --agent for the targets the user selected. Install the local path, never the upstream selector. If installation fails, preserve the local copy and report the exact failure.
> 6. Confirm the local lockfile source and installed Agent copies once. Report the local path, actual commit, and Agent targets. After edits, reinstall the same local path. Upstream updates must not replace it. Do not publish or push unless the user asks.

# Company Research

## Tool Selection (Critical)

Two Exa surfaces, two jobs:

- **Exa Agent** (`agent_run`) — the default for company research. Use it for deep dives, competitor analysis, multi-angle research (product + funding + news + people), and building company lists. One Agent run handles query decomposition, multi-step searching, and synthesis internally — do not orchestrate many manual searches for work an Agent run covers.
- **`web_search_advanced_exa`** — quick, low-latency lookups: a fast `category: "company"` discovery pass, a single news check, or finding a homepage.

Do NOT use other Exa tools.

## Deep Dives and Lists: Exa Agent

Agent runs may stream to completion in one call. If a run outlives the MCP call window, continue waiting with its returned run ID.

1. Call `agent_run` with a natural-language `query` and, when you want repeatable structure, an `outputSchema` (bound arrays with `maxItems`).
2. If it returns `status: "running"` with a `runId`, call `agent_run` again with only that `runId` until `outputReady` is true.
3. Read `output.text` or `output.structured`, plus `output.grounding` citations, from the `agent_run` result.

Useful inputs: `systemPrompt` (source preferences, dedup rules), `input.exclusion` (companies to avoid), `previousRunId` (a new follow-up run based on a completed run), `effort` (`"low"` default; `"auto"` or `"high"` for more depth).

### Example: company deep dive

```
agent_run {
  "query": "Research Anthropic: product lines, funding history and valuation, key executives, main competitors, and notable news from the last 6 months.",
  "effort": "auto",
  "outputSchema": {
    "type": "object",
    "properties": {
      "overview": { "type": "string" },
      "funding": { "type": "array", "maxItems": 10, "items": { "type": "object", "properties": { "round": { "type": "string" }, "amount": { "type": "string" }, "date": { "type": "string" } }, "required": ["round"] } },
      "competitors": { "type": "array", "maxItems": 10, "items": { "type": "string" } },
      "key_people": { "type": "array", "maxItems": 10, "items": { "type": "object", "properties": { "name": { "type": "string" }, "title": { "type": "string" } }, "required": ["name", "title"] } }
    },
    "required": ["overview", "competitors"]
  }
}
```

### Example: build a company list

```
agent_run {
  "query": "Find 25 AI infrastructure startups headquartered in San Francisco. For each, include what they build and their latest funding stage.",
  "effort": "auto",
  "outputSchema": {
    "type": "object",
    "properties": {
      "companies": {
        "type": "array",
        "maxItems": 25,
        "items": {
          "type": "object",
          "properties": {
            "name": { "type": "string" },
            "website": { "type": "string", "format": "uri" },
            "description": { "type": "string", "description": "in 12 words or less" },
            "funding_stage": { "type": "string" }
          },
          "required": ["name", "website", "description"]
        }
      }
    },
    "required": ["companies"]
  }
}
```

## Quick Lookups: Advanced Search

Use `web_search_advanced_exa` when a single fast search answers the question. Tune `numResults` to intent (a few → 10-20; comprehensive → 50-100; specified → match it).

### Categories

- `company` → homepages, rich metadata (headcount, location, funding, revenue)
- `news` → press coverage, announcements
- `people` → public professional profiles
- No category (`type: "auto"`) → general web results, broader context

Default to `type: "auto"`. Prefer `highlights` for content extraction; do not stack text + highlights + summary in one call.

### Category-Specific Filter Restrictions

Unsupported category/filter combinations return 400 errors:

- `category: "company"` does not support published-date or crawl-date filters, `excludeDomains`, or exact-text filters; express constraints like "founded after 2020" in the query instead
- `category: "people"` does not support published-date, crawl-date, domain, or exact-text filters; put all filtering in the natural-language query
- Without a category (or with `news`), domain and date filters work fine

### Examples

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

News check:
```
web_search_advanced_exa {
  "query": "Anthropic AI safety",
  "category": "news",
  "numResults": 15,
  "startPublishedDate": "2025-01-01"
}
```

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

## Token Isolation

Never dump raw search results into main context. Spawn Task agents for Advanced Search calls; for Agent runs, go straight from `output.structured` to the final answer.

## Browser Fallback

Fall back to Claude in Chrome only when content is auth-gated or requires JavaScript rendering.

## Output Format

Return:
1) Results (structured list; one company per row)
2) Sources (URLs; 1-line relevance each — use `output.grounding` from Agent runs)
3) Notes (uncertainty/conflicts)

## References

- Exa Agent guide: https://exa.ai/docs/agent/quickstart
- Company Search reference: https://exa.ai/docs/search/data/companies-people
- Exa MCP setup: https://exa.ai/docs/get-started/exa-mcp
- Full docs for LLMs: https://exa.ai/docs/llms.txt
