fix(cli): scaffold Valibot PORT as v.number()

[yamcodes]

Summary

• arkenv init with Valibot already imports @arkenv/standard/valibot, which binds @valibot/to-json-schema so v.number() coerces env strings.
• The generated PORT field still used v.pipe(v.string(), v.transform(Number), …), which fights that wrapper. Scaffold PORT as a numeric schema instead (same range/default as before).
• The examples/with-valibot example on v1 already used the wrapper + v.number(); this aligns the CLI with it.

Test plan

• pnpm --filter arkenv exec vitest run src/features/scaffold/validators.test.ts
• Confirm a Valibot arkenv init env file imports @arkenv/standard/valibot and has PORT: v.optional(v.pipe(v.number(), …), 3000)
• PORT=3000 boots without a manual transform

Made with Cursor

[changeset-bot]

🦋 Changeset detected

Latest commit: 7114998

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
arkenv	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[pkg-pr-new]

Open in StackBlitz

@arkenv/agent-plugin

npm i https://pkg.pr.new/yamcodes/arkenv/@arkenv/agent-plugin@1663

arkenv

npm i https://pkg.pr.new/yamcodes/arkenv@1663

@arkenv/build

npm i https://pkg.pr.new/yamcodes/arkenv/@arkenv/build@1663

@arkenv/bun-plugin

npm i https://pkg.pr.new/yamcodes/arkenv/@arkenv/bun-plugin@1663

@arkenv/core

npm i https://pkg.pr.new/yamcodes/arkenv/@arkenv/core@1663

@arkenv/fumadocs-ui

npm i https://pkg.pr.new/yamcodes/arkenv/@arkenv/fumadocs-ui@1663

@arkenv/nextjs

npm i https://pkg.pr.new/yamcodes/arkenv/@arkenv/nextjs@1663

@arkenv/nuxt

npm i https://pkg.pr.new/yamcodes/arkenv/@arkenv/nuxt@1663

@arkenv/standard

npm i https://pkg.pr.new/yamcodes/arkenv/@arkenv/standard@1663

@arkenv/vite-plugin

npm i https://pkg.pr.new/yamcodes/arkenv/@arkenv/vite-plugin@1663

commit: 7114998

[pullfrog]

ℹ️ No critical issues — one minor suggestion inline.

Reviewed changes

• Valibot scaffold PORT pipeline — packages/arkenv/src/features/scaffold/validators/dialects/valibot.ts now generates PORT: v.optional(v.pipe(v.number(), v.integer(), v.minValue(1), v.maxValue(65535)), 3000), dropping the v.string() + v.transform(Number) front-end in both the server field and defaultSimpleSchemaFields paths. Same range and default as before.
• Scaffold template assertion — validators.test.ts updated to pin the new numeric pipeline.
• Loader test fixture — jiti-schema-loader.adapter.test.ts updated from v.fallback(v.pipe(v.string(), v.transform(Number)), 3000) to v.optional(v.number(), 3000) for the "defaulted keys" case.
• Changeset — adds a patch changeset for arkenv describing the change.

I verified the fix behaviorally against @arkenv/standard/valibot (built from source): the new schema coerces PORT=3000 to a number, falls back to 3000 when unset, and rejects abc, out-of-range, and fractional values. The old scaffold was genuinely broken — with PORT unset it threw Expected string but received 3000 because Valibot routes the v.optional default through the pipe, so v.string() rejected the numeric default. Both affected test files pass locally.

  ^{｜ Fix all ➔ ｜ Fix 👍s ➔ ｜ View workflow run ｜ Using DeepSeek Flash (free via Pullfrog for OSS) ｜ 𝕏}

[pullfrog]

✅ No new issues found.

Reviewed changes

• Added a runtime regression pin in @arkenv/standard — the new commit boots the exact CLI PORT schema (v.optional(v.pipe(v.number(), v.integer(), v.minValue(1), v.maxValue(65535)), 3000)) with an empty env (expects default 3000) and with PORT="3000" (expects coerced number 3000), closing out the prior review's minor suggestion and the two still-unchecked boxes in the test plan. I confirmed its regression-catchy: a scratch pin using the old string-transform schema still throws on the empty-env path, so this test would have caught the original default-crash. All 48 CLI scaffold/loader tests and the 5 subpaths.test.ts tests pass against a fresh build.

No concerns with the incremental delta. The behavioral fix itself was already validated in the prior review.

  ^{｜ View workflow run ｜ Using DeepSeek Flash (free via Pullfrog for OSS) ｜ 𝕏}