How to implement Kanban tickets
A solo SaaS developer uses a custom MCP to let Roo retrieve and implement Kanban tickets. A separate QA agent returns failed work or routes it to human review.
news.ycombinator.comSuggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Retrieve [ticket] from [Kanban board] through [MCP integration] and implement its requirements in [repository]. Have a separate QA agent assess the result against the ticket. Return failed work with findings, or move passing work to [human review stage].
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source poster / publisher
- guyfromfargo
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The solo developer reports that AI completed about 90% of a full rewrite for a micro SaaS with around 50 paying customers, and estimates a week's coding can be completed in a day.
- Human involvement
- The workflow explicitly includes human review after agent QA; the source does not describe how that review is conducted.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Write a custom MCP integration to retrieve Kanban tickets.
- Have Roo pull tickets and begin implementing their requirements.
- Send the implementation through a separate QA agent.
A solo developer describes connecting Roo to a Kanban board through a custom MCP integration. They use the workflow on a real micro SaaS with approximately 50 paying customers, while working nights and weekends.
The developer first wrote the MCP integration to retrieve tickets. Roo then pulls those tickets and starts implementing them. A separate agent performs QA on the implementation. Depending on that assessment, it either sends the ticket back for more work or advances it to human review. The source describes a handoff to human review, but does not explain the checks performed there.
The author reports completing a full rewrite in which AI handled about 90% of the project. They also estimate that the setup lets them complete approximately a week's worth of coding in a day. These are the developer's own reported estimates; the comment supplies no independent measurements or breakdown of the remaining work.
The workflow was still being refined: the author says they had barely begun optimizing it. The concrete arrangement combines ticket retrieval, implementation, a separate QA decision and a human-review destination.
Who it's for: Solo developers with a ticket backlog who want automated implementation and QA while keeping a human review stage.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · guyfromfargo
Frequently asked questions
How did Roo receive work from the Kanban board?
The developer wrote a custom MCP integration to retrieve tickets. Roo pulls down those tickets and begins implementing them.
What happens after the implementation agent finishes?
Another agent performs QA, then either returns the ticket for more work or moves it to human review. The post does not describe the human review checks.
What results does the developer report?
The developer says AI completed about 90% of a full rewrite and estimates that the workflow delivers about a week's coding in a day. Both figures are self-reported.
More details
Remaining steps
- Return tickets for more work or move them to human review.
About this account
Roo retrieves and implements tickets, while another agent decides whether to return the work or advance it to human review.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=43086251. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: guyfromfargo on Hacker News
Source published
Added to Agents In Use
Source checked
Tools and topics
Roo · Custom MCP integration · Kanban board · QA agent
Kanban · MCP · Agent QA · SaaS
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.