Research companies, competitors, funding, news, leadership, and market context with Exa Agent and advanced search. Use when researching companies, competitor analysis, market research, or building company lists. Writes findings the caller can fold into brand/competitors.md or brand/landscape.md. Prefer this over ad-hoc web search for company deep dives. Distinct from competitive-intel (full brand competitive file methodology) - this skill is the Exa research engine those foundation skills should call.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
Research companies, competitors, funding, news, leadership, and market context with Exa Agent and advanced search. Use when researching companies, competitor analysis, market research, or building company lists. Writes findings the caller can fold into brand/competitors.md or brand/landscape.md. Prefer this over ad-hoc web search for company deep dives. Distinct from competitive-intel (full brand competitive file methodology) - this skill is the Exa research engine those foundation skills should call.
allowed-tools
["Bash","Read","Write","Skill"]
author
Exa Labs (ported by mktg)
license
MIT
user-invocable
true
metadata
{"openclaw":{"emoji":"๐ข"}}
On Activation
Read brand/positioning.md, brand/competitors.md, and brand/audience.md if present. Ground queries in the brand's category and known competitors. All optional.
Confirm Exa MCP Agent tools or EXA_API_KEY. If missing, stop with the install hint from Prerequisites / mktg doctor.
Default to Exa Agent for deep dives and lists; use advanced search only for quick single lookups.
When /cmo or a research agent owns the brand write, return structured findings + sources - do not silently overwrite brand/competitors.md unless the user asked to update brand memory.
Company Research
mktg runtime note
Prefer Exa MCP when available (tools: web_search_exa, web_search_advanced_exa, web_fetch_exa, agent_run).
If MCP Agent tools use the older create/wait/get names (agent_create_run, agent_wait_for_run, agent_get_run_output), use those equivalently.
Without MCP, call the HTTP API with x-api-key: $EXA_API_KEY (POST https://api.exa.ai/search, /contents, /agent).
Firecrawl remains the path for deep scrape of a known URL after Exa discovery.
Tool Selection (Critical)
Two Exa surfaces, two jobs:
Exa Agent (agent_run, or legacy agent_create_run / agent_wait_for_run / agent_get_run_output) - 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 are async: create the run, wait for it, then read the output.
agent_create_run with a natural-language query and, when you want repeatable structure, an (bound arrays with ). Returns an ID.
outputSchema
maxItems
agent_run_...
agent_wait_for_run until the run is completed (call again if still running).
agent_get_run_output - read output.text or output.structured, plus output.grounding citations.
Useful inputs: systemPrompt (source preferences, dedup rules), input.exclusion (companies to avoid), previousRunId (follow-up runs), effort ("auto" default; "high" for hard research).
agent_create_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 โ 10-20; comprehensive โ 50-100; specified โ match it).
Categories
company โ homepages, rich metadata (headcount, location, funding, revenue)
news โ press coverage, announcements
people โ public professional profiles
No category (type: "auto") โ general web results, broader context
Default to type: "auto". Prefer highlights for content extraction; do not stack text + highlights + summary in one call.
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
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:
Results (structured list; one company per row)
Sources (URLs; 1-line relevance each - use output.grounding from Agent runs)
