Redmine

Hey all,

I think this is my first post here, although I have used Redmine on and off at several companies over the years.

One recurring problem was getting Redmine accepted by users without adding commercial plugins. Redmine itself was always solid, but I generally needed paid additions to provide the portfolio overview and interactive planning experience people expected.

I have therefore started building a set of forever-free GPL plugins that aim to improve the standard Redmine experience without replacing its existing concepts or turning it into a different product.

The first two are:

Redmine Pulse
https://github.com/pljeroen/redmine_pulse

A portfolio and project-health dashboard that ranks visible projects and shows which ones need attention, together with the main reason why.

Redmine Pulse Plan
https://github.com/pljeroen/redmine_pulse_plan

An interactive planning layer for the existing Issues and Gantt views, including dependency-based scheduling, critical-path calculation, drag-and-drop editing, baselines, working calendars, and resource-levelling proposals.

The intention is to fit into the existing Redmine ecosystem, use native permissions and data wherever possible, keep plugin-owned data isolated and removable, and avoid telemetry, licensing checks, or external dependencies.

There is one unusual aspect that I want to disclose openly: these plugins were developed through an agentic engineering process.

I am a system and project engineer, not a fluent Ruby or Rails developer. I can slowly read and understand most of the code, but I could not have written these plugins conventionally. The implementation was produced by Opus 4.8 and GPT-5.5 following a process I designed around specifications, tests, independent review stages, compatibility checks, and evidence written to disk. The process does not accept an LLM saying that something works as evidence that it does.

That distinction matters to me: I did not personally write the code, but neither did I simply ask an AI to generate a plugin and publish the result. I engineered and enforced the process that produced it.

What I cannot provide myself is experienced Redmine plugin-developer judgement.

The process currently declares both plugins ready, but I would genuinely appreciate someone experienced taking a look and telling me where that judgement is wrong. In particular, feedback on Redmine conventions, permissions, host integration, migrations, upgrade safety, maintainability, and anything likely to create problems for other installations would be valuable.

I would also like to understand how the community feels about contributions produced this way. If AI-assisted contributions are unwanted regardless of their tests, documentation, and implementation quality, I will respect that and simply keep hosting them independently. If the method itself is not disqualifying, I would much rather have the work judged on what it actually does and how safely it does it.

The intention is not to pretend that I have suddenly become a senior software engineer. It is to contribute useful improvements to a project I have relied on for years, using capabilities that were previously unavailable to me.

All feedback is welcome, including the conclusion that the process or its output is inadequate. I would rather discover that now, refine the process, and repair these plugins than continue producing more of the wrong thing.