Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
Carl-Fredrik Arvidson avatar

Sapling

  • Updated May 17, 2026
  • cfarvidson/sapling

sapling is a Claude Code plugin and MCP dev workbench that stores knowledge and a typed work queue in Postgres so developers can enqueue and autonomously run coding-agent tasks through structured projects.

About

sapling is a Claude Code marketplace plugin wiring an AI-native MCP dev workbench with a Postgres-backed knowledge store and typed work queue. Developers reach for it when they want agents to pull structured tasks, run autonomously through plan-code-review cycles, and track projects to a verified Definition of Done instead of ad-hoc prompts. Slash commands enqueue work, manage projects and schedules, answer awaiting_input questions, and inspect queue health. A local runner daemon claims tasks and spawns agent commands, with an optional TUI dashboard for live supervision.

  • Postgres-backed MCP dev workbench exposing 49 tools plus /sapling slash commands
  • Typed work queue with plan, code, review, project, and schedule lifecycle management
  • sapling-runner polls the queue and spawns coding agents up to configurable concurrency
  • Projects auto-enqueue scoping, per-service plans, code, reviews, and DoD verification loops
  • make tui provides tmux-based live dashboard with per-agent windows and queue controls

Sapling by the numbers

  • Data as of Jul 10, 2026 (Skillselion catalog sync)
/plugin marketplace add cfarvidson/sapling

Add your badge

Show developers this marketplace is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Last updatedMay 17, 2026
Repositorycfarvidson/sapling

How do I turn ad-hoc agent prompts into a durable queue with projects, reviews, Definition-of-Done verification, and autonomous runner execution?

Queue and autonomously run coding-agent tasks via Postgres-backed MCP workbench with projects, plans, reviews, and a live TUI dashboard.

Who is it for?

Developers running local autonomous agent workflows who want Postgres-backed task queues, project orchestration, and MCP-exposed slash commands.

Skip if: Quick one-off Claude Code sessions without Docker, Postgres, or appetite for operating a local MCP server and runner daemon.

What you get

Claimed and completed work items, project status transitions, artifacts in the knowledge store, and optional TUI visibility into running agents.

Plugins in this marketplace

1 plugin - install individually after you add the marketplace.

Recommended Marketplaces

FAQ

What do I need for the quickstart?

Copy .env.example to .env, run make up for Postgres and mcp-server, then install the plugin with /plugin marketplace add cfarvidson/Sapling.

How does autonomous execution work?

Sapling-runner polls the MCP queue, spawns agent_command subprocesses up to max_concurrent, and agents self-claim via claim_next_work.

What is a Sapling project?

A project drives one intent across services with auto-enqueued scoping, plans, code, reviews, and a final Definition-of-Done verifier.

AI Agentsagentsautomation

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.