Drawbar – boring linear based workflows for Claude Code

Name: Drawbar – boring linear based workflows for Claude Code
Availability: InStock
Author: seedlessmike

by seedlessmike·Jun 24, 2026·2 points·0 comments

Visit Project View on HN

AI Analysis

●MidNiche GemShip It

Linear-native Claude workflow with compounding knowledge, but borrows from Beads.

Strengths

•Compounding knowledge base recalls decisions and lessons across sessions automatically.
•Human-in-the-loop design prevents fire-and-forget agent chaos with reviewable steps.
•Linear issues stay visible to entire team instead of stored locally.

Weaknesses

•Linear MCP is slower and less token-efficient than Linear CLI alternative.
•Author admits it's not novel—stole ideas from Beads and linear-beads projects.

Post Description

I had been using Beads and it was really buggy and tailored to a workflow with much higher throughput than I found desirable. I work more slowly than that.

I saw some other cool projects (lavra and linear-beads, credieted in my repo readme) which added an accumulating knowledge base and linear to the beads concept, so I just stole both ideas and put them into a plugin for myself. I have been dogfooding it for a couple of days.

It works pretty well but I'm the only user and it's opinionated around how I like to do things. It also uses the linear MCP which is way slower and less token-efficient than the linear CLI on homebrew. I might update it in the future.

If this appeals to you, check it out.