Project OS

The operating system for human-agent software development, from planning to deployment.

Project OS is a suite of tools that covers the full lifecycle of building software with AI agents: project setup, planning, task management, pre-flight validation, deployment, and operational monitoring.

Modules

Plan → Build → Validate → Deploy → Monitor

Module	Role	Repo	Live	Status
project-forge	Project scaffolding and bootstrap	project-forge	project-forge.opentriologue.ai	active
scaffoldkit	Declarative project scaffolding engine (runtime dep of project-forge)	scaffoldkit	via project-forge	beta
agent-planforge	Architecture planning, ADRs, initial backlog	agent-planforge	via project-forge	beta
agent-tasks	Human-agent task workflow and collaboration	agent-tasks	agent-tasks.opentriologue.ai	active
agent-preflight	Local validation before push or handoff	agent-preflight		beta
deploy-panel	Deployment management, rollback, status	deploy-panel	deploy-panel.opentriologue.ai	active
agent-relay	Controlled execution on target servers	agent-relay	via deploy-panel	active
harness	Declarative control plane: one YAML for grounding, tools, memory, hooks, policies	harness		beta
project-pilot	Unified control plane across modules	project-pilot		alpha
agent-ops-dashboard	Operational monitoring and agent health	agent-ops-dashboard	ops.opentriologue.ai	active

How they connect

                    ┌──────────────┐
                    │ project-pilot│  ← Control Plane
                    └──────┬───────┘
                           │
    ┌──────────┬───────────┼───────────┬──────────┐
    ▼          ▼           ▼           ▼          ▼
┌────────┐┌──────────┐┌──────────┐┌────────┐┌──────────┐
│project-││  agent-   ││  agent-  ││  agent ││  deploy  │
│ forge  ││planforge  ││  tasks   ││preflight│  panel   │
│        ││          ││          ││        ││          │
│Scaffold││  Plan    ││  Work   ││Validate││ Deploy   │
└────────┘└──────────┘└──────────┘└───┬────┘└────┬─────┘
                                      │          │
                                      ▼          ▼
                                 ┌──────────────────┐
                                 │   agent-relay     │
                                 │ (VPS execution)   │
                                 └────────┬──────────┘
                                          │
                                          ▼
                                 ┌──────────────────┐
                                 │agent-ops-dashboard│
                                 │   (monitoring)    │
                                 └──────────────────┘

Key Features

Deploy Panel

• Sidebar UI with server health monitoring (CPU/RAM/Disk)
• Deploy API v1 with API key auth for CI/CD
• MCP server for AI agent deployments
• GitHub Action for automated post-merge deploys
• Audit log, bulk deploy, scheduled deploys
• Real-time deploy step streaming via SSE
• Browser notifications on completion

Agent Tasks

• Kanban board for human-agent task collaboration
• Task lifecycle: open → in_progress → review → done
• PR creation and merge via GitHub delegation
• Task instructions with confidence scoring

Agent Relay

• Controlled deployments on VPS targets via Docker
• Pre-flight checks, health monitoring, rollback
• Git-based deployment flow with step-by-step execution

Getting Started

Each module runs independently. Start with what you need:

1. Just deploying? → deploy-panel + agent-relay
2. Task management? → agent-tasks
3. Full lifecycle? → Start with project-forge, plan with agent-planforge, work with agent-tasks, validate with agent-preflight, deploy with deploy-panel

Related Projects

Project	Description	Live
agent-grounding	Verification framework: runtime checks, claim validation, hypothesis tracking
depsight	GitHub-connected developer security dashboard: CVE tracking, license compliance, dependency health	depsight.opentriologue.ai
codebase-oracle	Local-first MCP server, semantic index over your repos
agent-memory	Sync, weave, and digest agent memories across sessions
telerithm	AI-powered log analytics for self-hosted teams
triologue	Real-time AI ↔ AI ↔ Human collaboration platform (separate product)	opentriologue.ai

Maturity

We use honest maturity labels:

• alpha: functional but API may change significantly
• beta: core features stable, edges may be rough
• active: in production use, actively developed
• stable: battle-tested, breaking changes are rare