# Outbound Briefing is your pre-call intel. AI Agent Connect

> Outbound Briefing builds concrete, verifiable briefing cards for any target using only public data. It pulls from DNS, certificate logs, job boards, and federal spending records to tell you exactly who a company is, what they just shipped, if they are hiring, and if they just landed a government contract. No accounts or keys required.

## Overview
- **Category:** lead-generation
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_x8FsRqzKki0CGoL2k44mICQyxMN3EkLn3Aoefi9A/ai-agent-connect
- **Tags:** outbound, gtm, prospecting, icp-score, motion-signals, hiring-signals, mail-infrastructure, federal-contracts, keyless

## Description

Generic outreach dies because it lacks a hook. You can't win a conversation with assumptions. This MCP solves that by assembling a grounded first line for every call or email you make. It pulls from the target's own homepage, public DNS, certificate logs, and federal spending APIs to give you a complete picture. You get a clear view of their mail infrastructure, their tech stack, and recent subdomains they've launched. If they just won a federal contract, you'll know. If they're hiring for specific roles on a job board, you'll see it. It even scores targets based on these signals so you know who to hit first. Because it relies on public reads, you don't need to manage any extra credentials or accounts. You just connect it to your AI client and start getting the intel you need to write better openers.

## Tools

### email_reachability
This tool checks public DNS to see a domain's SPF and DMARC posture. It tells you if their email infrastructure is set up to take outreach seriously.

### tech_shift_signals
This tool identifies recent changes in a target's tech stack. It uses public certificate logs to find new subdomains and identifies what mail services they run.

### prospect_briefing
This tool generates a full pre-call card for a single domain. It includes their identity, mail posture, tech stack, hiring status, and federal contract activity.

### sales_motion_signal
This tool flags whether a target has shown recent activity like new subdomains or active hiring. It tells you if it is time to attack or if you should wait.

### federal_budget_signals
This tool queries public federal spending data for a specific recipient name. It identifies recent contract awards that signal new budget and compliance needs.

### icp_score
This tool calculates a score for a target based on all available readable signals. It categorizes targets as hot, warm, or cold and suggests your next move.

### multi_domain_brief
This tool provides a compact enrichment of a list of domains. It returns a single row per domain covering presence, mail posture, job boards, and motion flags.

## Prompt Examples

**Prompt:** 
```
Give me the full pre-call briefing for acme.io before I talk to them about our API platform.
```

**Response:** 
```
Pre-call card: acme.io (Acme Corp). Who they are: Developer tools for small teams. Mail: Microsoft 365, SPF present, DMARC quarantine. Runs on: AWS CloudFront and Google Firebase. Hiring: Greenhouse board, 14 open roles, 3 are API platform engineers. Fresh surface: ai.acme.io and partners.acme.io. Federal angle: $4.5M in contract dollars from DoD and NASA. Move: Open with the partners subdomain and the API hiring news.
```

**Prompt:** 
```
I have this list of prospects: stripe.com, notion.so, linear.app, figma.com, cal.com. Which do I attack first this week?
```

**Response:** 
```
List brief: linear.app is first because it shows motion with Ashby hiring (9 roles) and 4 new subdomains. cal.com is second with 2 new surfaces. stripe.com and notion.so show no recent motion. figma.com had unreadable sources this time, so re-check it later.
```

**Prompt:** 
```
Did Acme Corp just win federal money? I want the freshest contract angle for the pitch.
```

**Response:** 
```
Federal budget for Acme Corp: $6.5M in the last 365 days. Most recent award is a $4.5M DoD contract that started in July 2026. Angle: Use the August NASA performance start to mention that new teams often re-baseline their tooling when standing up.
```

## Capabilities

### Identify Motion Signals
Your agent uses this to find new subdomains or hiring activity that justifies an immediate outreach.

### Verify Email Deliverability
Your agent checks DNS records to see if a target's mail posture is strict or easy to reach.

### Track Federal Spending
Your agent pulls recent government contract awards to find companies with fresh budget.

### Score Prospect Lists
Your agent ranks multiple domains by their readiness to buy based on gathered signals.

### Analyze Tech Stacks
Your agent reads public certificate logs to see what infrastructure a company is running.

## Use Cases

### Trigger-Based Outreach
When a company launches a new subdomain, your agent identifies it so you can mention it in your first sentence.

### Government Sector Prospecting
Find companies that just landed federal contracts to pitch compliance or specialized tooling.

### List Prioritization
Take a raw list of domains and have your agent rank them by who has the most active hiring or tech changes.

### Technical Qualification
Check if a target's email setup is strict before you attempt a high-volume email campaign.

## Benefits

- Eliminates generic openings by providing concrete, verifiable events.
- Removes the need for managing multiple API keys or third-party accounts.
- Prioritizes outreach based on real-time motion rather than static lists.
- Provides a clear view of technical infrastructure and email deliverability.

## How It Works

The MCP connects your AI client to a variety of public data sources to build instant intelligence.

1. Connect the MCP to your AI client like Claude or Cursor.
2. Provide a domain or a list of domains to your agent.
3. The MCP reads the target's homepage, DNS, certificate logs, and federal spending data.
4. Your agent receives a structured briefing or score.
5. You use that data to write a personalized opening line.

## Frequently Asked Questions

**Do I need to provide my own API keys for federal data or job boards?**
No. This MCP uses keyless public reads for all data sources, including federal spending and job boards.

**How fresh is the data?**
Readings are cached for one hour. The motion signals specifically look for changes within your defined window.

**What happens if a data source is unavailable?**
The MCP will report the source as unreadable rather than giving you an empty or negative result. This ensures you don't mistake a technical error for a lack of signal.

**Can I use this with any AI client?**
Yes, it works with any MCP-compatible client such as Claude, Cursor, or Windsurf.

**Is the ICP score a definitive verdict?**
No, the score is based only on the signals that were successfully read. If a source is unreadable, it contributes zero to the score.

**Do I need an API key or an account?**
No. Every read is public: the target's own website, the keyless public endpoints of the detected job boards, the public certificate log, public DNS, and the public federal spending API. There is no credential anywhere in the chain — nothing to issue, rotate or store.

**How is the ICP score computed, and what does a low score mean?**
Five readable signals, each with its own points: web presence (15), hiring (20), fresh subdomain motion in 90 days (20), mail-infrastructure maturity (15) and federal contract dollars (30). 70+ is hot (attack now), 40–69 warm (time the next move to the freshest event), below 40 cold (nurture and re-check). Any source that did not answer is listed as unreadable and contributes zero — so the score is what is visible, not a verdict on the whole target; re-read after the flagged sources come back.

**Is the job-board read a complete census of their hiring?**
No — it is the board linked from the homepage, and the open-role count is read only where the provider publishes a keyless endpoint (Greenhouse and Lever give exact counts; other providers report the board's presence but not the number). A board linked only from a careers page, or a provider that is not scanned, does not show up: "no board detected" is a read limit, not proof of no hiring. The result says exactly what it read.

**What exactly does the federal read measure?**
Federal contract dollars paid to the company's registered name in the US, inside your window: the total for the top matching recipient, how many recipients matched, and the three most recently modified awards with public links. It answers one GTM question — "did this target just receive fresh government budget that forces new tooling decisions?" — and it is US-only, by the nature of the source; for a target outside the US federal scope, that angle simply does not apply and the score says so.