AI Agent Orchestration Platform

Forked and extended an open-source AI company orchestration platform into a custom deployment with multi-agent hierarchies, role-based heartbeat systems, and MCP tool integration. Integrated the BMAD methodology with 45+ workflow skills spanning the entire SDLC — from market research and PRD creation through architecture design, story development, and QA. Built a tiered skill variant system that reduces token consumption by 50-87% across agent fleets, and developed custom plugins for MCP bridging, authentication, and workspace initialization. Agents operate autonomously with heartbeat-driven execution, git-tracked workspaces per issue, and budget enforcement.