Projects
Recent and active projects from Jeremy Longshore at Intent Solutions — what's been shipping in the last few weeks, with star counts and live links.
What’s been shipping lately. Star counts and pushed dates pulled live from GitHub; pages on this site link out to the canonical repo or live URL.
Flagship — open source
claude-code-plugins-plus-skills · 2,100+★
The marketplace and ecosystem hub. 425 plugins, 2,810 skills, 200 agents for Claude Code — installable via the ccpi CLI package manager. Public marketplace at tonsofskills.com; source at github.com/jeremylongshore/claude-code-plugins-plus-skills.
This is the through-line for most of the other work — every plugin and skill listed below ends up in this hub. The ccpi package manager handles install, validation, and upgrade.
@intentsolutions/audit-harness
Deterministic test-quality gates published as a single dev dependency. Hash-pinning, escape-scan, CRAP, architecture, bias, and Gherkin lint — runnable as pnpm exec audit-harness <cmd> in any repo. Companion to the /audit-tests and /implement-tests skills (the canonical Intent Solutions testing SOP).
Recently shipped
guidewire-mcp-for-claude
Carrier-native MCP servers + governance harness for Guidewire estates (PolicyCenter, ClaimCenter, BillingCenter). v0.1.0 shipped in May 2026 — six foundation packages, five carrier-vocabulary tools, full blueprint published the same day. Anthropic-first; will publish to the claude-code-plugins-plus marketplace.
plugins-nixtla · 6★
Nixtla time-series forecasting plugins — StatsForecast, MLForecast, NeuralForecast — exposed as Claude Code skills with proper agent affordances. Built for an active Nixtla engagement. GitHub
claude-code-slack-channel · 17★
Two-way Slack bridge for Claude Code via Socket Mode + MCP stdio. Hash-chained tamper-evident audit journal, per-thread session isolation, policy-gated MCP tools, permission relay with Block Kit buttons, five-layer prompt-injection defense. Three runtimes (Bun / Node.js / Docker), TypeScript strict. GitHub
j-rig-skill-binary-eval
Release-quality evaluation harness and rollout gate for Claude Skills. Pushed daily; the eval framework that shipped 10 epics in a single day during March 2026. GitHub
intent-blueprint-docs · 27★
Enterprise AI documentation generator — 22 templates covering PRD, architecture, tasks, risk management. Works in Claude Code and Cursor, zero-setup. GitHub
iam-bobs-brain · 21★
General-purpose enterprise orchestrator on Google ADK + Vertex AI Agent Engine. Multi-agent system with risk tiers (R0–R4), policy gates, evidence bundles, and Mission Spec v1 workflows. GitHub
gastown-viewer-intent · 29★
Real-time agent observability dashboard. Beads task graph, agent activity stream, and system health rendered as either TUI or React web interface. Go + React. GitHub
Production hosting
Braves Booth Intelligence
Real-time broadcast operations dashboard for Atlanta Braves radio. Microservices, LLM narrative generation with a 4-tier fallback chain, live probability gauges. v1.0.0 currently live at scorecardecho.com.
Intent Solutions VPS
A single Contabo VDS hosts five containerized stacks behind one Caddy ingress: Plane (project management), Twenty (CRM), Umami (analytics), ntfy (notifications), and the Braves Booth stack. Watchtower auto-updates the off-the-shelf services; per-repo CI/CD handles the bespoke ones. The migration runbook is open: intentsolutions-vps-runbook.
Partner deliverable portal
Single Hugo site at partners.intentsolutions.io with sections per partner engagement, deployed via Tailscale OIDC + force-command SSH from a GitHub Actions workflow. ~22 seconds end-to-end including smoke checks.
Tooling and skills
excel-analyst-pro · 26★
Financial modeling toolkit for Claude Code with auto-invoked skills. Build DCF, LBO, variance, and pivot tables from natural language. GitHub
iam-git-with-intent · 10★
AI-powered git workflow automation — smart commit messages, changelogs, PR descriptions, and release notes generated from diff analysis. GitHub
contributing-clanker
The /contribute Claude Code skill plus the OSS contribution workspace it operates on. Keeps AI-assisted upstream contributions honest before they reach maintainers — runs deterministic gates against any AI-proposed PR. GitHub
qmd-team-intent-kb · 10★
Governed team memory platform for Claude Code, powered by QMD. Teams build, query, and maintain shared knowledge bases with access controls and audit trails. GitHub
Research and exploration
irsb · 1★
On-chain guardrails for AI agents — EIP-7702 spend limits, cryptographic execution receipts, automated dispute resolution. The principle: no AI agent should hold unguarded keys. GitHub · hub on this site
moat · 1★
Verified Agent Capabilities Marketplace — MCP-first trust, policy, and execution layer for agent-to-agent interactions. GitHub
intentional-cognition-os
Local-first, remote-capable knowledge OS. “Compile knowledge for the machine. Distill understanding for the human.” Active exploration. GitHub
cad-ai-agent · 3★
Drawing intelligence platform for AEC professionals. Upload a DXF, PDF, or DWG; describe what you need in plain English; get structured edits, compliance reports, quantity takeoffs, RFIs, and drawing summaries — all without the AI ever touching the original file. GitHub
This site
Start AI Tools is the daily journal — built on Hugo, deployed via Netlify, rendered through the Dispatch design system, and published into automatically by a 7 AM cron that classifies each day’s work into a tier and writes the post end-to-end. 241 posts published as of this writing, growing daily without manual intervention. Source: github.com/jeremylongshore/startaitools.com.
Most repos are MIT or Intent Solutions Proprietary. Need help implementing any of these? Let’s talk.