Set up an email agent that works Redmine tasks

A developer wired custom agents to take job requests by email and let one of them handle Redmine tasks inside a locked-down sandbox.

Done with Custom agent· Intermediate· Nov 2026

What they did
They built a set of custom agents on top of RubyLLM and usually fed them instructions through email. One agent could read a Redmine project board, take on tasks from it, and carry out the work. For development, it also had access to a sandbox container without sudo rights.
How it went
The setup was run with a locally hosted Qwen model, but no performance results were shared. The only concrete safeguard mentioned was the non-sudo sandbox, which restricted what the agent could alter while working.
Tips
If you try this, keep the agent’s filesystem and permissions limited so it can only touch what you intend. Add a review step before any task updates or email replies go out.

Try it yourself with n8n

Create a workflow for [dedicated email address] that watches for messages from [approved sender list], extracts the requested work, and turns it into a task in [Redmine/project board]. Then let an AI agent work on the task in [sandboxed environment] with no sudo access, and draft an email back describing what it changed. Stop before sending any reply or modifying any task status, and ask me to approve those actions first. Show me a test run using a sample email before I turn it on.

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Read the original on news.ycombinator.com ↗

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.