Agent / Productivity Tools

Most repetitive work is too small to justify a platform and too annoying to keep doing by hand: reformatting the same kind of document every week, drafting the same kind of checklist from scratch each time, moving the same information between two places that don’t talk to each other. Buying software for any one of these is disproportionate, and building a bespoke internal tool for it is usually somebody’s unpaid side project that never quite gets finished. The task stays exactly as annoying as it was, because nothing addresses problems this specific and this small. None of it is important enough on its own to get budget, a project plan or a place on anyone’s roadmap, so it just gets absorbed into however much longer the working day needs to be.

Each tool in this set starts as an internal fix for one specific, narrow task — something that kept costing more time than it should inside DarkHorseOne itself — and gets published once it’s clear the problem wasn’t unique to us. That internal-first path means every tool that reaches this set has already done its job once, for real work, before it was written up as something anyone else could use. One example still active in the set is an agent that drafts an onboarding checklist directly from a job description, turning a document that already exists into a starting checklist rather than a blank page.

Tools are also consolidated when it makes sense: a standalone note-capture tool was folded into the main toolset rather than kept running as a separate product once its job was better served from inside it. Shared workspaces for small teams — letting more than one person work against the same tool — are in early development rather than finished; some of what’s here today is a working experiment, not a settled feature. Which of these two things happens to any given tool — consolidation, or continued standalone development — depends on how it’s actually used once people other than the original team try it, not on a plan fixed in advance.

This set suits individuals and small teams who have a specific, narrow task that keeps costing more time than it should, and who are comfortable using a tool that started as somebody else’s internal fix rather than a purpose-built commercial product. It suits people who want a focused tool that does one job well rather than a platform they have to configure to do several, and who don’t mind being an early user of something still finding its final shape in exchange for a tool built around a real, specific annoyance rather than a generic use case nobody actually has.

It does not suit anyone who needs a tool guaranteed to keep working unchanged: these begin as experiments, and the intended lifecycle includes some being replaced and some being retired once they’ve served their purpose or a better version takes over — the note-capture tool folded into the main toolset is one example of that happening already. It does not suit a team that wants one comprehensive platform to standardise on, since the whole point of the set is narrow, single-purpose tools rather than a unified suite. It does not suit a procurement process that requires a committed release schedule to sign off on, either — none is published, because the direction each tool takes depends on what turns out to be useful once it is out. And because the tools are labelled Experimental rather than a stable release, it does not suit anyone building a process they can’t afford to rebuild if the specific tool underneath it changes or disappears.

These tools are free to use, with no separate paid tier attached to them. To start, get in touch through the contact page — DarkHorseOne’s engineering team reads every message directly and usually replies within two working days. There is no public catalogue to browse yet, and because the set is still forming, what’s available changes over time — the contact page is also the way to find out what currently exists.