How it works

Four steps from a brief to a skill your agents can load, with a fixed price at the start and the evidence it worked at the end.

1. 01

   Send your brief

   What your agents should get better at, the questions the skill has to handle, and how you would know an answer was right. One flat fee at checkout. The test questions are how the work gets marked, so they are worth more thought than the brief.

2. 02

   Our researchers read the field

   We research your topic area in detail and review everything you sent (videos, documents, reports), selecting and interpreting the material under a fixed lens: what is claimed, whether it holds, and what would falsify it.

3. 03

   We write the skill

   Written to trigger on the right work and to say only what your agents do not already know. Every claim in it is checked back against the source it came from.

4. 04

   We test it and hand it over

   Run against the questions you set, with the skill loaded and without it. You get the skill and the side-by-side answers, so you can see what it changed before you deploy it.

What we build from

Our researchers explore your topic in detail. Not a search and a summary. We find the people who actually know the subject, review their talks, lectures, papers and books, and read them properly under a fixed lens: what is claimed, whether it holds, and what would falsify it.

That research alone is enough to build you a skill. But, if you send us reference material as well, it gets read alongside: videos and recordings (send a link and we fetch and transcribe them), documents, reports, slide decks, internal handbooks, style guides, past deliverables. Anything that describes how the work is actually done where you are makes the skill more specific to you, but nothing here is a prerequisite.

Not only code, and not only video: recorded talks and lectures, books, industry reports and white papers, your own internal documents, and published web material. If the thinking exists somewhere in prose or speech, it can become a skill.

Yours, and it runs anywhere

What we hand over is a folder in the open Agent Skills format. Anthropic published the specification in December 2025 and released it as an open standard, and the rest of the industry adopted it: one SKILL.md file of Markdown with a short block of metadata at the top, saying what the skill is for and when an agent should reach for it, alongside any scripts, templates or reference files it needs.

You get the whole folder. The skill itself, the trigger text tuned so it fires on your kind of work and stays quiet on the rest, any reference material it depends on, and the test questions with the answers it produced. It goes into your repository and it is yours to read, edit and version like anything else you own.

The format is shared; tools differ a little in where they expect skills to live and how they scope them to a project. So we hand yours over with the exact path for each tool your team runs, which is all it takes. There is nothing to integrate.