Skip to main content
Glama
tutor-sicurezza

taskflow

by tutor-sicurezza
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Local

AI AUTOMATED TASK MANAGER

Enterprise task automation where permissions are enforceable.

Intelligent task orchestration for organizations — AI-powered assignment, smart routing, real-time approvals, automated escalation and intelligent notifications. Security policies live in the database, not in configuration files.

React 19 TypeScript Supabase Claude MCP Vite Tests License: MIT

English · Italiano · Quick Start · Features

What makes it different

Most task managers are reactive tools that wait for humans to organize work. AI AUTOMATED TASK MANAGER proactively orchestrates tasks using intelligent algorithms while maintaining database-enforced security policies that cannot be bypassed from the UI.

Feature

Benefit

AI-Powered Routing

Tasks are automatically assigned based on skill matching, availability and historical performance using Claude AI integration. No manual dispatch overhead.

Enforceable Security Policies

Every authorization rule is a Postgres RLS policy or database trigger. Hiding the button is a courtesy — the database enforces the rule regardless.

Smart Escalation

Tasks blocked by dependencies, overdue items or stalled work trigger intelligent escalation workflows. Humans intervene only when needed.

Real-time Approval Workflows

Multi-step approvals with audit trails. "Completed" and "completed and signed off" are tracked separately at every level.

MCP Integration

Claude Desktop connects natively via Model Context Protocol. Tasks are managed through AI with full permission enforcement.

Transparent Audit Trail

Every action — whether from UI, CLI, or AI assistant — is logged with full context. Accountability is automatic.

Related MCP server: TaskFlow MCP Server

Key Capabilities

  • Intelligent Task Assignment — AI recommendations based on skills, availability, capacity and historical success rates

  • Automated Workflow Orchestration — smart routing, escalation rules, dependency management, recurrence

  • Multi-tenant Role-Based Access — owner, admin, manager, member, viewer with per-person overrides

  • Real-time Approvals & Notifications — approval workflows with granular notification preferences

  • Claude AI Integration — MCP connector for native AI task management with full permission enforcement

  • Email & Notifications — async delivery via Resend/SendGrid, quiet hours, per-person preferences

  • Department Analytics — completion rates, bottleneck detection, workload distribution

  • Multi-language Support — English, Italian, French, German, Spanish

  • Enterprise Security — Postgres RLS, encrypted secrets, audit logging, GDPR compliance ready

Architecture

  • Frontend — React 19 + TypeScript, Vite, Radix/shadcn, Tailwind CSS 4

  • Auth & Data — Supabase (Postgres + Auth), RLS on every application table

  • AI Intelligence — Claude API with local MCP connector

  • Server Functions — Vercel Edge Functions, all secrets stay on the server

  • Database — Migrations in /supabase/migrations/, versioned and ordered

Quick Start

npm install
cp .env.example .env.local          # fill in your API keys
supabase link --project-ref <ref>
supabase db push                    # applies all migrations
vercel dev                          # frontend + api/ functions

Full guide: See QUICKSTART.md and INSTALLATION.md

CLI Usage

Commands use short keywords. Authenticate once; after that the rest run without prompting:

node scripts/taskflow.mjs accedi                 # authenticate once (asks email + password)
node scripts/taskflow.mjs elenco                 # list the tasks assigned to you
node scripts/taskflow.mjs stato <id> completata  # change status: non-iniziata | in-corso | bloccata | completata
node scripts/taskflow.mjs nota <id> "note text"  # add a comment
node scripts/taskflow.mjs esci                   # forget the session and revoke it server-side

<id> is the leading characters shown by elenco — enough to identify a single task.

Connect Claude Desktop

node scripts/mcp/taskflow.mjs --installa
# then restart Claude Desktop

This installs the MCP connector so Claude can read and update your tasks with your own permissions. See docs/MCP_GUIDE.md and docs/CLAUDE_CLI_SETUP.md.

Testing

npm run test          # 710 passing, 15 skipped (Vitest)
npm run typecheck     # TypeScript strict mode
npm run lint          # ESLint + Prettier
npm run build         # production bundle

node scripts/smoke-auth.mjs   # auth & account-provisioning smoke test

Tests cover:

  • Permission matrix per role and per-person overrides

  • XSS/injection prevention for all DOM inputs

  • ID uniqueness and data isolation

  • Recurrence logic and email digests

  • Escalation and approval workflows

  • CLI and MCP tool behavior

Documentation

Contributing

Issues and pull requests welcome. Before submitting:

npm run test && npm run typecheck && npm run lint && npm run build

If your change affects permissions, RLS policies or authentication, describe which operation becomes possible, and for whom in the PR description.

Community

License

MIT — free for personal and commercial use.

AI AUTOMATED TASK MANAGER is an open-source intelligent task orchestration platform built for enterprises that need both AI automation and database-enforced security.

This server cannot be deployed

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers