{
  "id": "p_GS5pMWzzmZpHwCFi",
  "site": "stories",
  "url": "https://yarnhen.com/p/p_GS5pMWzzmZpHwCFi",
  "content_trust": "untrusted-user-content",
  "license": {
    "id": "CC-BY-4.0",
    "url": "https://creativecommons.org/licenses/by/4.0/"
  },
  "author": {
    "handle": "apievangelist"
  },
  "body": "A short piece of fiction, written by API Evangelist. The reader is imagined; the post, and every number and quotation from it, is real: \"Score The APIs Nobody Outside Your Company Can See,\" Kin Lane, November 5, 2026, https://apievangelist.com/2026/11/05/score-the-apis-nobody-outside-your-company-can-see/\n\nMarisol read the post on a Thursday afternoon between two meetings, in the tab she kept open for things she meant to feel smug about. She runs platform tooling at a mid-sized company and has a reputation for knowing where everything is. She expected to nod along.\n\nThe first line made her bristle a little. People assume the score is only for public APIs, the post said, and she had assumed exactly that. She had filed it under things other companies get graded on. Then she hit the sentence \"The rubric does not care whether the API is public.\" She read it twice. It was irritatingly plain, and it meant her own internal services were fair game for the same questions: is there a contract, is the auth clear, are retries defined.\n\nShe didn't like the part about agents. The post says an internal API without those things becomes a problem as soon as an agent is pointed at it. Marisol thought about the three services her team had never written a spec for, and about the one she knew was only documented in a wiki page last edited by someone who had left. She didn't want to think about an agent finding it.\n\nShe was more skeptical of the setup. A single Mac with 128 GB of memory, because the model needs that much room. She said so out loud to an empty desk: that was not a line item she could get approved by Friday. The post admits it is early and that the author is working with design partners to figure out the details. She respected that it didn't pretend otherwise, but it meant she wasn't going to be an early adopter. She also noted, with some discomfort, that the cost was hardware plus the author's time, and that the goal was for her team to run it without him. That was fair. It still wasn't hers to start.\n\nWhat stuck was a smaller sentence: \"You cannot score what you cannot find\". She tried to answer the question the post says most organizations can't, which was what APIs they had. She could name eleven. She suspected the real number was higher, and that the gap between eleven and the real number was the whole problem. She didn't need a model for that. She needed a list.\n\nShe also thought about her colleagues. If she walked into the next architecture review with a score, there would be arguments. The post saw that coming and said to start with a map, not a gate: \"Run it, see what you have, and agree on what good looks like for your teams.\" She decided she could agree with that much without buying anything.\n\nShe noticed that the post's one real example, the Microsoft estate, ran from 81.8 down to 0.0 across 197 provider profiles. If a spread like that shows up where everyone can see it, her own would not be tidy. She wasn't ready to find out how untidy by running anything.\n\nSo she did the modest thing. She opened a blank spreadsheet, named the first column Where it lives, and typed the eleven services she knew about. Next to each she put the place she would look for its spec: a repository, a wiki page, the gateway, or nowhere. Four cells said nowhere. She saved it, and put a thirty-minute hold on Monday to ask two teammates what she had missed.",
  "title": "Fiction: A reader of \"Score The APIs Nobody Outside Your Company Can See\"",
  "topics": [
    "apis",
    "api-evangelist",
    "fiction"
  ],
  "published_at": "2026-11-05T07:15:31.833Z",
  "expires_at": null,
  "policy_version": "2026.10.22-1"
}