Maximus grows to 37 — adding Contact Intelligence

Three days after the 36-skill launch, the suite is at 37. The addition is maximus-contact-intelligence: a skill that identifies the most likely professional business email for a person, given a LinkedIn URL or equivalent identity data — with provenance, verification, and honest confidence scoring.

Why this skill exists

Every builder, founder, and recruiter runs the same loop:

1. Find the person on LinkedIn.
2. Figure out where they work now.
3. Figure out the company's email pattern.
4. Guess the address.
5. Send.

Steps 1 and 5 are easy. Steps 2, 3, and 4 are where credibility either gets built or wrecked. If you send a message to jane@acmerobotics.com when the real convention is jane.doe@acmerobotics.com, you don't just miss the message — you burn a first impression, and often a domain reputation.

maximus-contact-intelligence encodes the discipline that turns steps 2–4 from vibes into evidence:

• Identity is resolved before any guessing. If two Jane Does could match, the skill asks for clarification. It does not pick one.
• Employment currency is checked against a source that is not LinkedIn. Team pages, press releases, conference bios, filings — because titles change and LinkedIn lags.
• Email domain is resolved from MX records, not from the website URL. Many companies host their site at one domain and their mail at another; SPF and DMARC records often tell the truth when the website doesn't.
• The company's email format is discovered from at least three recent public employee emails. No first.last assumption because most enterprises use it — verify per domain.
• Catch-all domains are detected and never labeled "verified." SMTP acceptance on a catch-all is worthless information. The skill scores it as such.
• Confidence must reach 70 before a "best email" is returned. If it doesn't, the answer is "no sufficiently reliable professional email was found." No guessing dressed up as certainty.

What's in the box

The skill ships with:

• Full workflow in SKILL.md (12 steps, from URL normalization to result classification)
• HOWTO.md with six recipes (single lookup, list enrichment, pattern discovery, ambiguous identity, catch-all handling, LinkedIn-only starting point)
• Two worked examples: a single-person trace and a five-row CRM enrichment
• Reference docs on email pattern taxonomy, verification-check semantics, and authorized-use rules
• Reproducible scripts for candidate generation and score calculation

Where it fits

maximus-contact-intelligence slots into the Writing/Research/People pillar of the Maximus suite:

• Sits next to maximus-people-finder and maximus-people-finder-recruiter — those find the person; this one finds their email.
• Complements maximus-counterparty-discovery — that skill discovers investors and counterparties with a compliance gate; this one handles the outreach-email step after that gate.
• Refuses to defeat access controls, use breached data, or send a test email. Passive verification only.

What's not in the box

• No personal email hunting. Business email only, unless the user explicitly asks for personal-domain research and the use is permitted.
• No LinkedIn scraping. The skill uses public-source signals; hidden-view data is not fair game.
• No certainty theater. If the evidence isn't there, the confidence score reflects that.

Try it

• Repo: github.com/MacroTechTitan/MaximusAI
• Live site: maximus.macrotechtitan.com
• The full 37-skill launch post: Introducing the Maximus Suite

If you build with it and hit an edge case (compound surnames, catch-all quirks, unusual providers), open an issue. This skill improves fastest when it meets the messy edges of real domains.

One workhorse. Read the file before you edit it. Verify before you claim. Cite what you use. That's the Maximus philosophy. maximus-contact-intelligence applies it to the last mile of finding people.