Research account
Research one company before a meeting and hand back a briefing — what it does, what it publicly says is hard right now, and who it names as competition, each line traceable to where it came from.
Install research-account and follow its SKILL.md, starting from https://api.getcargo.io/INSTALL.md$npx skills add getcargohq/gtm-skills/research-accountNo account needed to install. A new Cargo account starts with 100 free credits, no card.
Say this to your agent
I have a discovery call with Fabrikam tomorrow — we sell data-pipeline monitoring. Brief me.
Illustrative output #
Fictional records, to show the shape of what comes back.
Fabrikam (fabrikam.example) — Industrial IoT, ~1,100 staff, Munich [linkedin]
What they do: sensor analytics for factory lines, sold to OEMs [site]
Priorities, last 12 months:
- Moving plant telemetry to a single cloud warehouse [press, 2026-03]
- "Fewer, better dashboards" for plant managers [CEO talk, 2026-06]
Hiring: 6 data engineers, 1 Head of Data Reliability (posted 9 days) [jobs]
Likely: pipeline breakage is a live pain — rests on the reliability hire
Unknown: current monitoring vendor; no public statement
Ask them:
1. What broke the last time a plant's data went missing?
2. Who owns the warehouse migration, and when is it done?
3. What will the Head of Data Reliability be measured on?
One briefing for 1.175 credits (0.5 + 2 × 0.025 + 0.125 + 0.5): every line sourced, one marked Likely:, one marked unknown.
What it costs #
| Action | Credits |
|---|---|
linkedin.enrichCompanyFromDomain | 0.5 |
parallel.extract | 0.025 |
parallel.createTask | 0.125 |
theirStack.searchJobs | 0.5 |
Never run this across a full list on the first attempt. Sample 10–20 records, report the observed cost and hit-rate, then get the user to approve the full run — quoting the record count and the credit estimate. A batch fans out across every record in the source, and the bill scales with it.
One account is about 1.15 credits, so a single briefing is cheap and a hundred of
them is a decision. Say the total before running a list. parallel.extract bills
per URL, so two pages is 0.05 and ten is 0.25.
Worth knowing #
- Most of this reads public material, so a private company with a thin web presence returns little. That is a real answer: report the gap rather than padding the page.
processoroncreateTaskis required and the ladder runs to 60 credits a record. Pinlite. Escalate tobase(0.25) on a specific hard target, never across a list.theirStack.searchJobsfilters on--data, not on--records. Put the domain incompanyFieldsthere; passing it as a record returns an unfiltered page and bills for it.- Ask for the domain, not the company name. Every call here keys off it.
- For a renewal, drop steps 2 and 3 and keep the baseline and the roles. The account team already knows the story; what they want is what changed.