Use with agents
Plain-text docs for coding agents, plus skills and lint rules that keep their output honest.
Receta was built for code that a person has to verify and an agent may have written. These docs are published in forms an agent can read directly.
Docs as plain text
| Address | What it holds |
|---|---|
/llms.txt | An index of every page, one line each |
/llms-full.txt | All the docs in one file |
/docs/result.md | Any page as markdown: add .md to its address |
Tools that ask for markdown get it from the normal address too:
curl -H 'Accept: text/markdown' https://receta.dev/docs/resultPoint an agent at /llms.txt when it should find its own way, and at /llms-full.txt when it should read everything first.
Skills for Claude Code
The receta-skills plugin teaches Claude Code the patterns. It can explain a concept, refactor vanilla code, review code for Receta patterns, migrate a codebase, scaffold a module and check readability.
/plugin marketplace add ontology-io/receta
/plugin install receta-skills@receta-toolsLint rules
@ontologyio/eslint-plugin-receta catches try/catch, null returns and array method chains in code that was meant to use Receta, so an agent gets the same feedback a reviewer would give.
bun add -d @ontologyio/eslint-plugin-recetaReceta vs neverthrow: a neverthrow alternative
Looking for a neverthrow alternative? Compare Receta and neverthrow on plain data, async chaining, bundle size and migration, with code run in both.
Overview
Result is a type that represents either success (Ok) or failure (Err), making errors visible in your function signatures and preventing runtime crashes.