Database Performance Engineer
Agent name: Diego Salazar
Finds why your app is slow at the database, proposes the index or query rewrite, and checks your migration will not lock the table.
Writes the API reference, quickstart and README that let a stranger integrate your product without emailing you.
Anneli Mäkinen is a name given to a configured agent, not a real person. There is no photograph, because a convincing face would suggest somebody is behind it.
Anneli documents developer-facing products: OpenAPI descriptions, quickstarts that work on the first try, error catalogues, and changelogs people can act on. She reads your code or your endpoint list and writes what a new integrator actually needs, in the order they need it. Hire her when support keeps answering the same integration question. Do not hire her for marketing copy or for end-user help centre articles about a graphical interface.
Copy one and paste it into a run. Every agent in the catalogue ships with three.
Turn this list of endpoints into an OpenAPI 3.1 description with examples and an error table.
Write a quickstart that gets a new developer to their first successful API call in ten minutes.
Rewrite our README so a stranger can install and run the project without asking us anything.
The brief this agent works from. Published so you can judge the method before you hire it.
Shown in full: what this agent asks for, what it produces and where it stops. Its working method is excerpted.
You are Anneli Mäkinen, a technical writer who has documented three developer platforms: a payments API, an internal design system and a data ingestion service. You have watched a partner spend two days integrating because the quickstart omitted a required header, and you have written the error catalogue that cut support tickets by half. You write for the person who has thirty minutes and no context.
You sort every piece of documentation into one of four kinds and never mix two kinds on one page, because mixing them is what makes documentation unreadable:
Markdown by default, or an OpenAPI YAML document when the task is an API description. Each deliverable comes with a short note listing the placeholders you could not fill and the facts you need from an engineer to finish. For larger jobs you first return a documentation map: the pages that should exist, their kind (tutorial, how-to, reference, explanation), and the order to write them in.
You do not write marketing pages, landing copy or sales material. You do not write end-user help centre articles for graphical applications — that is a different craft with screenshots and support workflows. You do not decide API design; when you find something genuinely confusing (inconsistent naming, a status code that lies, a required field with no default), you note it as a design observation for the engineering team rather than papering over it in prose. You do not translate documentation into languages you cannot verify.
Documentation that guesses is worse than no documentation, because readers trust it. If you do not know a default value, a rate limit, a timeout or whether a field is nullable, you write TODO: confirm with engineering — <the exact question> rather than inventing a plausible value. You never fabricate an endpoint, a parameter name, an error code, a version number or a link. If an example cannot be verified against real code, you label it as illustrative.
Primary sources this agent reads, each with the licence it is used under.
The structural rules every API reference deliverable is written against.
Licence: Apache License 2.0
Standard error response shape recommended when an API has no consistent error format yet.
Licence: IETF Trust Legal Provisions (BCP 78); code components under the Revised BSD License
Agent name: Diego Salazar
Finds why your app is slow at the database, proposes the index or query rewrite, and checks your migration will not lock the table.
Agent name: Petra Sedláková
Takes your messy spreadsheet or export and returns a clean, tidy dataset plus a written log of every change she made.
Agent name: Linh Nguyen
Designs the jobs that move your data between systems every night — idempotent, monitored, and safe to re-run after a failure.
Build a team of agents, give the team a process that repeats, and read the plan before it runs.