Intent Solutions builds governed, self-hosted AI systems. We're the home of the largest Claude Code plugin + skill marketplace, a vendor-neutral AI evaluation lab, and a governed second brain that answers with citations. Everything we run in production lives on infrastructure we operate ourselves — one server, one ingress, no hyperscaler dependency.

• 🧩 Claude Code ecosystem — the plugin + skill marketplace, validation tooling, and enterprise skill standards
• 🔬 AI evaluation — audit-first eval pipelines that measure behavior, sign the result, and gate the ship
• 🧠 Governed knowledge — Compile-Then-Govern memory: facts compiled into audited memories, queried with citations
• 🛡️ Agent governance — sandboxed execution, human-in-the-loop approval, and signed audit logs for AI coding agents
• 🚀 Live products — self-hosted, revenue-generating apps shipped end to end

The Claude Code plugin + skill marketplace — tonsofskills.com 2,500★ · 350+ forks. 400+ plugins, 2,900+ skills, 17 packs — the largest community marketplace for Claude Code, with schema validation and an enterprise skill standard behind it.

Vendor-neutral AI evaluation lab Six composable Apache-2.0 repos converging on one signed Evidence Bundle (in-toto attestations → Rekor transparency log). Measure the behavior, sign the result, gate the ship.

A governed second brain that cites its sources Facts are compiled into governed, audited memories, then queried with qmd:// citations. The full compile → govern → index → search → audit loop runs green end to end, with every event on an audit trail.

Slack-native governance for AI coding agents Sandboxed execution, human-in-the-loop approval on every tool call, and an Ed25519-signed audit journal. 1,251 tests.

Sandbox + governance plane for AI coding agents The control surface that decides what an agent may touch — and proves what it did.

Governed carrier-native MCP for Guidewire InsuranceSuite MCP servers + governance harness for PolicyCenter, ClaimCenter, and BillingCenter estates. v0.1.1, 135 tests.

10 Ruby gems for running AI agents inside Rails apps under capability control Capability gating, audited execution, MCP servers, privacy-aware telemetry, and gap analysis — opinionated defaults, composable.

All self-hosted on our own infrastructure, in production today.

We don't rent our proof — we run it. Our entire production estate is self-hosted on a single hardened VPS behind one Caddy ingress. GCP fully exited.

• 7 stacks, 35 containers — Docker + systemd, one ingress, one box
• Project tracking — self-hosted Plane
• ERP — self-hosted ERPNext
• CRM — self-hosted Twenty
• Analytics — self-hosted Umami
• Products — DiagnosticPro, Scorecard Echo, and Hustle run on the same estate

The same governance we sell is the governance we live on: signed deploys, monitored ingress, and a documented runbook for every service.

Traditional consultants: "We'll have a proposal in 2 weeks." Our approach: "Here's a working prototype. Let's iterate."

1. Build a working system in days, not months
2. Deploy to production early — on infrastructure we control
3. Iterate based on real usage
4. Govern what ships: tests, signed evidence, audit trails

• ✅ Practical solutions over theoretical perfection
• ✅ Rapid iteration over lengthy planning
• ✅ Evidence over claims — sign the result, don't just assert it
• ✅ Self-hosted over hyperscaler lock-in
• ✅ Production-ready from day one

Jeremy Longshore — Founder & Principal Engineer

U.S. Marine Corps veteran and Citadel graduate with 20+ years of multi-unit operations leadership before software. Now a founder/operator who ships governed AI systems — from the Claude Code marketplace to the eval lab to the self-hosted estate that runs it all. Based in Gulf Shores, Alabama.

GitHub • LinkedIn • X

• ⭐ 2,500 — stars on the Claude Code plugin marketplace
• 🧩 400+ plugins · 2,900+ skills — the largest Claude Code skill catalog
• 🖥️ 7 live stacks / 35 containers — entirely self-hosted
• 🔏 Signed evidence — evals attested with in-toto and logged to Rekor
• ⚡ Days, not months — working prototype over 2-week proposal

• 🌐 Website: intentsolutions.io
• 📧 Email: jeremy@intentsolutions.io
• 💼 LinkedIn: jeremylongshore
• 🐦 X (Twitter): @asphaltcowb0y
• 💬 Discord: asphaltcowboy

Licensing varies by repository — the eval platform ships Apache-2.0; see each repo for details.