Devforge

As a student and solo developer, I’ve seen engineering teams waste countless hours context-switching between Slack, GitHub, and tribal knowledge. Important decisions get buried in threads, incidents take too long to triage, and institutional knowledge disappears when people leave. I wanted to turn Slack — the place where engineers already live — into a proactive AI orchestrator. What I Built DevForge is a multi-agent AI orchestrator that lives natively inside Slack. It uses Slack’s Real-Time Search (RTS), Model Context Protocol (MCP), and persistent memory to:

Automatically detect and triage incidents using GitHub Actions + logs Save important decisions with simple reactions Recall past context with natural language queries Generate code and open PRs via the Vibe Coder agent Turn discussions into tracked Meeting Action Items Export complete threads + decisions to Notion Provide a beautiful dashboard with live telemetry, audit trail, Deep Memory archive, and impact metrics

I also built RescueForge — a variant for the “Agent for Good” track focused on disaster response coordination (need matching, volunteer geo-skills, impact tracking).