A short piece of fiction, written by API Evangelist. The reader is imagined; the post, and every number and quotation from it, is real: "Migrating from Spectral to OpenLint: we're looking for beta testers," Kin Lane, October 29, 2026, https://openlint.org/blog/2026/10/29/migrating-from-spectral-beta-testers/
Marisol maintains the linting setup for a mid-sized company's API platform team. She found the post on a Thursday afternoon, between a failing pipeline and a meeting she was avoiding. The title made her wince, because she had Spectral wired into a lot of places and had never wanted to think about it.
She read the opening lines with her guard up. OpenLint 1.0.0 was on npm, the specification was published, the docs were live. All of that was true of plenty of projects she had ignored. What held her attention was the second paragraph. The post said the next step was moving over, and that the authors wanted beta testers to help prove the path. She noticed that it didn't claim the path was already proven.
She was warier of the section about API Commons switching. A project that runs its own tools on its own linter is not an independent witness, and she said so to the empty room. Still, the post gave a concrete claim she could weigh: the results had been checked against Spectral 6.16.1 on every fixture they had, and they matched. Every fixture they had was not the same as every fixture she had. She had a ruleset with three custom functions written by someone who had since left, and she wasn't sure anyone alive understood one of them.
That made the list of snags the useful part. She read it slowly. The first item hit her directly: "A function that imports @stoplight/spectral-core won’t load in OpenLint." She opened her repository and searched. Two of the three custom functions imported exactly that. She felt a small, irritated relief at learning it from a blog post rather than from a red pipeline at the worst moment.
The second snag bothered her more. Rulesets carrying extra properties, a title or a prompt for an AI agent, were rejected, and the authors were stripping them from published rulesets for now. Her team had added that kind of metadata on purpose, because people read those rulesets. Asking her to delete documentation so a tool would accept the file felt like the wrong direction. She didn't think it was a scandal, since the post said backwards compatibility was an open discussion, but she wasn't going to pretend she liked it.
The third was minor but familiar. A clean run with the JSON format prints an empty list plus a line of text, and she had a script that parsed that output. She knew it would break.
She didn't feel converted. She still wasn't sure why she should move at all, and the post didn't try very hard to persuade her. It asked for testers and made no promises. But she did like one line near the end: "Just knowing who’s trying it helps us decide what to fix first." It asked for something small.
She didn't touch the production pipeline. She made a branch, ran npx openlint lint against her least important ruleset, and left the custom functions alone. Then she opened Discussions on the project and started a thread saying she was testing, that she'd be coming back with the three snags from her own repository, and that the extra properties were the one she cared about most.
She closed the laptop and went to her meeting.