Skip to content
View rdmilly's full-sized avatar
  • RevenueFirst.AI
  • Daytona Beach, FL
  • 14:48 (UTC -08:00)

Block or report rdmilly

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
rdmilly/README.md

Hi, I'm Ryan

I ran a moving company for nine years. I started it with one truck and grew it to 25 people doing hundreds of moves a year around Portland, and I handled everything from dispatch and hiring to quotes and payroll.

Running it showed me where service businesses lose time and money, so now I build the automation that fixes that. I design the systems, write the specs and acceptance criteria, and direct AI coding agents to build them. Then I test and verify everything before it ships.

The biggest thing I've built is a self-hosted controller that lets AI agents like Claude safely use around 690 tools across four servers, with a build, test and security gate in front of every release. The build log is at gocontroller.app.

What's here: the memory platform that gives my AI tools long-term memory (helix-memory-platform) and the Chrome extension that feeds it (membrain-extension), the pipeline I use to find and qualify leads for my automation work (lead-sourcing-pipeline), and the MCP servers that connect Claude to the tools my businesses run on (business-mcp-servers). Those four are public copies of private working repos, with the secrets and server addresses taken out.

I also built a gate and blind-review loop that pushes the sites I build until they read as hand-made (website-refinement-loop).

Smaller pieces: a discovery scanner in Go (paving-agent), an MCP tool for git operations (git-agent), a health daemon for my servers (hostinger-watchdog), local push-to-talk dictation for Windows (whisper-ptt), my MCP gateway design (mcp-tool-flow), and the skill files I use to run Claude sessions (claude-skills).

I'm studying for CompTIA Security+ and looking for work in AI implementation, automation, solutions engineering or technical project management, remote or around Daytona and Orlando. LinkedIn

Pinned Loading

  1. mcp-tool-flow mcp-tool-flow Public

    Complete MCP Tool Flow Architecture - OAuth Gateway, Semantic Tool Filter, Name Bridge, Context Forge aggregation. 95% token reduction for Claude.ai integration.

    JavaScript 1

  2. paving-agent paving-agent Public

    The Paving Agent: single-command discovery scanner that maps a node, installs watchtower coverage at every discovered service, and opens a pre-loaded AI conversation with full system context. Part …

    Go

  3. business-mcp-servers business-mcp-servers Public

    MCP servers that let Claude work with the tools my businesses run on: DNS, secrets, invoicing, email, social, logs, knowledge bases. Public copy of a private working repo.

    Python

  4. helix-memory-platform helix-memory-platform Public

    Self-hosted memory for AI tools: captures conversations, extracts decisions, RAG over Postgres/ChromaDB/Neo4j, injects context into new chats. Public copy of a private working repo.

    Python

  5. lead-sourcing-pipeline lead-sourcing-pipeline Public

    Finds and qualifies small service businesses from public licensing data: crawl, enrich, score, push to outreach. Public copy of a private working repo.

    Python

  6. membrain-extension membrain-extension Public

    Chrome extension that gives AI chats persistent memory: captures, compresses and injects context across Claude, ChatGPT, Gemini and Perplexity. Public copy of a private working repo.

    JavaScript