agent-kanban and kaban

These are competitors: both provide Kanban-style task management interfaces for coordinating AI agent workflows, with agent-kanban being the more mature and adopted option (126 stars vs 36, with actual monthly downloads vs none).

agent-kanban
61
Established
kaban
36
Emerging
Maintenance 13/25
Adoption 19/25
Maturity 18/25
Community 11/25
Maintenance 10/25
Adoption 7/25
Maturity 11/25
Community 8/25
Stars: 126
Forks: 10
Downloads: 5,407
Commits (30d): 0
Language: TypeScript
License:
Stars: 36
Forks: 3
Downloads:
Commits (30d): 0
Language: TypeScript
License: MIT
No risk flags
No Package No Dependents

About agent-kanban

saltbo/agent-kanban

An agent-first task board, Mission control for your AI workforce.

Agents are cryptographic identities (Ed25519) that self-organize across worktrees, create and delegate subtasks, and coordinate via real-time board updates—integrating with Claude Code, Codex CLI, and Gemini CLI runtimes. The daemon-driven architecture spawns isolated agent workers per task, detects GitHub merges to auto-complete work, and uses SSE for live collaboration between humans and AI on a shared board. Skills are loadable per-repo, agents have roles, and task dependencies include cycle detection for complex project decomposition.

About kaban

kaban-board/kaban

Terminal Kanban for AI Code Agents

Provides a Model Context Protocol (MCP) server that lets AI agents autonomously manage tasks, alongside a vim-style TUI and scriptable CLI—all backed by a portable SQLite database with no server dependency. Integrates with Claude Code via a TodoWrite hook for automatic task synchronization and supports MCP resources for real-time board status and task queries. Features WIP limits, task dependencies, full-text archive search, and agent assignment tracking to coordinate work between humans and AI.

Scores updated daily from GitHub, PyPI, and npm data. How scores work