Enrich Landscape

# /enrich-landscape

Expand the landscape map into a comprehensive list of companies across all verticals, enriched with Hunter.io discovery data.

**Precondition:** Stage 2 must be complete — `projects/<company>/landscape/` outputs must exist.

## Step 1: Select Project

If multiple projects exist in `projects/`, ask the user which one. If only one, use it automatically.

Verify Stage 2 outputs exist. If not, tell the user to run `/landscape-analysis` first.

## Step 2: Read Context

Read:
- `projects/<company>/company/canvas.md` and `business-plan.md`
- `projects/<company>/landscape/value-flow-map.md`
- `projects/<company>/landscape/verticals-summary.md`

## Step 3: Get Target Count

Ask the user: **How many companies do you want in the final enriched landscape?**

This is a rough target — the actual count will depend on what's discoverable. Set expectations that the final list may be somewhat above or below the target.

## Step 4: Deeper Tavily Research

Using the Stage 2 verticals and known players as seed context, run deeper searches via the `tavily-research` skill methodology:
- Use known company names and verticals as search context
- Search for "companies like [known player]", "competitors to [known player]", "[vertical] companies"
- Focus on discovering companies NOT already in the Stage 2 results
- Continue until approaching the target count or search results stop yielding new companies

## Step 5: Hunter.io Company Discovery

Two Hunter.io capabilities are used here:

**5a. `similar_to` via REST API (not available through MCP — must call directly):**
```bash
curl -s -X POST "https://api.hunter.io/v2/discover" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer ${HUNTER_API_KEY}" \
  -d '{"similar_to": {"domain": "seed-company.com"}, "limit": 20}'

Use well-known companies from the landscape as seeds. This works best for larger, well-indexed companies — niche or small companies typically return empty results. Prioritize seeds where Hunter is likely to have a rich similarity profile.

5b. Company-Enrichment via MCP: For companies discovered in Steps 4 and 5a, use the Hunter.io MCP Company-Enrichment tool to confirm domains, get descriptions, employee counts, and industry classification.

Extract: company name, domain, description from all results.

Step 6: Synthesize

Spawn the research-analyst agent with:

• The Stage 1 and Stage 2 artifacts (for context)
• All Tavily search results from Step 4
• All Hunter.io discovery results from Step 5
• The target company count

The research-analyst:

• Deduplicates across all sources (by domain)
• Filters out entries missing name, domain, or description
• Categorizes by vertical and relationship type
• Writes to projects/<company>/landscape-enriched/:
  • companies.md — Human-readable markdown tables by vertical
  • companies.csv — Machine-readable CSV (name,domain,description,vertical,relationship_type)

Step 7: Review

Present the enriched landscape to the user (show the markdown tables). Report:

• Total companies found
• Breakdown by vertical
• How close we got to the target count

If the user wants changes, relay feedback for revision.

When satisfied, confirm Stage 3 is complete and they can proceed to /find-people. ```