[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78071YC5E7C7VF65DC7G9KM":3,"skill-same:01M78071YC5E7C7VF65DC7G9KM":45},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M78071YC5E7C7VF65DC7G9KM","effect-api-extractor-bases",null,"Use when API Extractor reports ae-forgotten-export for the anonymous base of an Effect class factory (Schema.Class, TaggedClass, TaggedError, Opaque, Context.Service) under the `effected package build` pipeline — and for the OTHER ae-*\u002Ftsdoc-* diagnostics a package build surfaces, ae-unresolved-link above all ({@link} selector rules for merged value+type names, overloaded functions (the parenthesized index selector), namespace-object members, inherited members, schema-declared Schema.Class fields, and cross-package symbols, where backticks are the only correct form), plus how to read issues.json without being fooled. The house policy for bases is to write the factory INLINE and suppress the synthesized X_base warning narrowly via effected.toml [package.build.meta.tsdoc] suppressWarnings (the default already carries { messageId ae-forgotten-export, pattern _base }) — no @public base const, no hand-written annotation. Yields a zero-warning issues.json with the base warnings in the suppressed bucket.","spencerbeggs\u002Feffected","spencerbeggs","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F109114?v=4","01M77ZTHB70K5Y2VVTRCFY50A5","C. Spencer Beggs","01M77ZTHCC41NXBT3X906K6QM1",8,"MIT","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fbuilds\u002Fcopilot\u002Fskills\u002Feffect-api-extractor-bases\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: effect-api-extractor-bases\ndescription: Use when API Extractor reports ae-forgotten-export for the anonymous base of an Effect class factory (Schema.Class, TaggedClass, TaggedError, Opaque, Context.Service) under the `effected package build` pipeline — and for the OTHER ae-*\u002Ftsdoc-* diagnostics a package build surfaces, ae-unresolved-link above all ({@link} selector rules for merged value+type names, overloaded functions (the parenthesized index selector), namespace-object members, inherited members, schema-declared Schema.Class fields, and cross-package symbols, where backticks are the only correct form), plus how to read issues.json without being fooled. The house policy for bases is to write the factory INLINE and suppress the synthesized X_base warning narrowly via effected.toml [package.build.meta.tsdoc] suppressWarnings (the default already carries { messageId ae-forgotten-export, pattern _base }) — no @public base const, no hand-written annotation. Yields a zero-warning issues.json with the base warnings in the suppressed bucket.\n---\n\n# API Extractor × Effect class factories\n\n`class X extends Schema.Class\u003CX>(\"X\")({...}) {}` (and `TaggedClass` \u002F\n`TaggedError` \u002F `Opaque` \u002F `Context.Service`) produces an anonymous\nheritage type. API Extractor reports it as `ae-forgotten-export` on a\nsynthesized `X_base` symbol — CI-fatal under `effected package build`.\n\n## The policy: inline factory + scoped `_base` suppression\n\nWrite the factory **inline** — no split-out base const, no hand-written\nannotation — and suppress the synthesized-base warning narrowly. The build's\ndefault `meta.tsdoc.suppressWarnings` is exactly that one rule, so a package\nwith no `effected.toml` already has it. A package that sets its own list\nreplaces the default and must restate the rule in its `effected.toml`:\n\n```toml\n[package.build.meta.tsdoc]\nsuppressWarnings = [{ messageId = \"ae-forgotten-export\", pattern = \"_base\" }]\n```\n\n```text\n\u002F\u002F src\u002FX.ts — inline, no base const, no annotation\n\u002F** ... @public *\u002F\nexport class SemVer extends Schema.Class\u003CSemVer>(\"SemVer\")({\n  major: nonNegativeInteger,\n  minor: nonNegativeInteger,\n  patch: nonNegativeInteger,\n}) {\n  \u002F** ... *\u002F\n  static readonly FromString = \u002F* ... *\u002F;\n}\n```\n\nThat yields `dist\u002Fprod\u002Fissues.json` = `{ warnings: 0, errors: 0, suppressed: N }`,\nwith the `SemVer_base` (and every other `*_base`) entry in the **`suppressed`**\nbucket.\n\n### Why this is principled, not a \"silence the warning\" cop-out\n\nThe emitted `.d.ts` proves the base is genuinely internal:\n\n```text\ndeclare const SemVer_base: Schema.Class\u003CSemVer, \u002F* full inlined shape *\u002F>;\ndeclare class SemVer extends SemVer_base { \u002F* full shape: fields, statics, methods *\u002F }\nexport { SemVer \u002F* , ... *\u002F };   \u002F\u002F \u003C- SemVer_base is NOT exported\n```\n\n`X_base` is a module-local `declare const` that never appears in the export\nlist, and the exported class carries its full shape. A consumer can neither\nname nor need `X_base`. The suppression is **narrow** (`pattern: \"_base\"`\nmatches only these synthesized symbols), **accounted for** (it lands in the\n`suppressed` bucket and the build log reports the count — it is logged, not\nsilenced), and **harmless** (`ae-forgotten-export` means \"a public API names a\ntype a consumer can't import\"; here nothing a consumer needs is lost). That is\ncategorically different from muting, say, an `ae-missing-release-tag` you don't\nfeel like fixing.\n\n### Validated across every factory kind\n\nConfirmed clean (`warnings: 0`, the `*_base` in `suppressed`) with `tsgo`\ngreen and **no hand-written annotations** on:\n\n- `Schema.Opaque`, `Schema.Class`, `TaggedClass`, `TaggedError`.\n  (Naming trap: there is no `Schema.asClass` — zero occurrences in\n  `Schema.ts`. To give a schema value a class identity you subclass it\n  directly: `class MyString extends Schema.String {}`.)\n- **Recursive `Schema.Class` + `Schema.suspend`** (a node whose field\n  references itself): the inline form has **no TS2506** and needs none of the\n  old `Schema.Schema\u003CSelf>` base-annotation gymnastics — that failure mode was\n  an artifact of splitting the annotated base const out of the heritage clause,\n  and it evaporates inline. The suspend callback still takes its normal\n  `(): Schema.Schema\u003CSelf> => Self` return type; that is ordinary recursive-schema\n  practice, not a base annotation.\n- **`Context.Service`** (`class R extends Context.Service\u003CR, Shape>()(\"id\") {}`):\n  same `R_base` suppression, no annotation.\n- **Named field-schema consts** (`major: nonNegativeInteger`): the const's type\n  **inlines into `X_base`** rather than emitting as `typeof nonNegativeInteger`,\n  so it does **not** forgotten-export and does **not** need `@public`. The old\n  \"schema helpers referenced by the annotation become `@public`\" cascade retires\n  with the annotation that caused it.\n\nSo the former ceremony — the `@public X_base` export, the mandatory explicit\nannotation, the recursive `Schema.Schema\u003CSelf>` special-casing, re-exporting\nevery base from `index.ts`, and marking field consts `@public` — is all gone.\nWrite the class the way Effect's own docs write it and add the one-line\nsuppression.\n\n## The suppression is `_base`-scoped ONLY — it does not cover method signatures\n\nThe binary release-tag rule still applies to **method and function\nparameter\u002Freturn types**: *anything a `@public` signature names must itself be\n`@public`*. An internal type on a public method signature is a **different\nsymbol** (not suffixed `_base`), so `pattern: \"_base\"` does **not** suppress it —\nby design, so genuine surface leaks are never masked. In the yaml port, one\ninternal `RawDiagnostic` parameter on a `@public`\n`YamlDiagnostic.fromRaw(raw: RawDiagnostic, …)` produced **12**\nforgotten-export warnings (the whole `internal\u002Fdiagnostics` module). These are\nreal and must be fixed, not suppressed:\n\n- **Inline a structural type** on the public signature so no internal symbol is\n  named — best for engine-internal record types that should not become public\n  surface:\n\n  ```text\n  static fromRaw(\n   raw: { readonly code: YamlErrorCode; readonly message: string; readonly offset: number; readonly length: number },\n   text: string,\n  ): YamlDiagnostic { … }\n  ```\n\n  (This is the **landed** form, not a hypothetical: `@effected\u002Fyaml`'s\n  `YamlDiagnostic.fromRaw` ships exactly this structural-inline signature — the\n  12-warning `RawDiagnostic` incident above is what motivated it.)\n\n- **Or tag the referenced type `@public`** and re-export it — only when it is\n  genuinely part of the API.\n\nPrefer the structural-inline form for anything under `src\u002Finternal\u002F`. Watch for\nthis on `X.fromRaw` \u002F `X.of` \u002F codec-adapter statics that bridge the internal\nengine to the public classes — exactly where an internal record type sneaks\nonto a `@public` signature.\n\n**Three escapes that look like they should work and do not.** Two build cycles\nwere burned re-discovering this, so it is worth stating outright:\n\n| Attempted escape | Result |\n| --- | --- |\n| Mark the *method* `@internal` and leave it on the `@public` class | **Still fails.** The gate follows the class's release tag; an `@internal` method of a `@public` class is still reachable rollup surface. |\n| Mark the *referenced type* `@internal` | **Still fails**, and differently — now you get `ae-incompatible-release-tags` instead. `@internal` never resolves an `ae-forgotten-export`; it only relabels the problem. |\n| Extract the shape to a named `type` alias and use that | **Still fails.** An alias is a named symbol like any other, so it forgotten-exports under its own name. Naming the thing is precisely what you cannot do. |\n\nOnly two things actually work: **inline the type structurally** (no name exists,\nso nothing can be forgotten), or **export it** and accept it as public API. There\nis no third option, and the effort spent looking for one is unrecoverable.\n\n### The positive counterpart: third-party generics are free\n\nThe mirror-image fear — that naming a *dependency's* generic type on a `@public`\nsignature will forgotten-export it — is unfounded. **API Extractor resolves\ndeclared-dependency third-party types as externals and emits real `import`\nstatements in the `.d.ts`**: no `ae-forgotten-export`, no inlining, no\nre-export ceremony. Confirmed at `@effected\u002Fgithub`'s first build with octokit's\n`Endpoints[R]` on a `@public` signature.\n\nThe condition is that the package be a **declared dependency** (it must be\nresolvable from the emitted `.d.ts`, which a `devDependency` or a transitive\npackage is not). Given that, write the precise third-party type rather than\nwidening to `unknown` or hand-copying its shape — the gate does not charge you\nfor it.\n\n### …nor does it cover a type in the class factory's GENERIC\n\nThe same gap bites one step earlier, in the **heritage clause itself**. An\nunexported interface passed as a *type argument* to the factory —\n\n```ts\nimport { Context } from \"effect\"\n\ninterface VersionCacheShape {\n  readonly current: string\n} \u002F\u002F NOT exported from index.ts\n\nexport class VersionCache extends Context.Service\u003CVersionCache, VersionCacheShape>()(\n \"@effected\u002Fsemver\u002FVersionCache\",\n) {}\n```\n\n— leaks as a plain `ae-forgotten-export` on **`VersionCacheShape`**, whose symbol\nname has no `_base` suffix, so `pattern: \"_base\"` does not match it.\n\n**Measured** on `@effected\u002Fsemver` by un-exporting exactly that interface and\nrebuilding:\n\n| | warnings | suppressed | code | `ciFatal` |\n| --- | --- | --- | --- | --- |\n| baseline | 0 | 12 | — | — |\n| interface un-exported | **1** | 12 (unchanged) | `ae-forgotten-export` | **`true`** |\n\n> `The symbol \"VersionCacheShape\" needs to be exported by the entry point index.d.ts`\n\nThe suppressed bucket stayed at 12 and every entry in it ends in `_base` — the\nnarrow suppression does exactly what it says and nothing more.\n\nThe fix is to **export the shape type** from the entrypoint (it is genuinely part\nof the API — it is the service's contract), not to widen the suppression. This is\nthe case a reviewer walks you into: nitting \"dedupe this shape into a named\ninterface\" is right on the merits, but the moment the name exists it must also be\nexported, or the gate goes red. Applies equally to `Schema.Class\u003CSelf, Fields>`\nand every other class factory that takes a type argument.\n\n## TSDoc `{@link}` traps\n\n**The one rule the cases below are instances of:**\n\n> `{@link}` resolves **only to symbols the entry point exports, under the name it\n> exports them by**. Everything else is a backtick span.\n\nSix ways a name fails that test:\n\n- **A schema-declared `Schema.Class` field** — the class owns it, but only\n  through the factory, which API Extractor does not see through.\n- **A module-local renamed at export** — `class Route` exported as\n  `export { Route as RestRoute }` resolves as `{@link RestRoute}` and **never**\n  as `{@link Route}`, no matter that `Route` is the name in the file you are\n  writing the comment in. Confirmed twice at the `@effected\u002Fgithub` build.\n- **A member of a const-only namespace object** — link through the exported\n  const (`{@link Walker.ascend}`), never the bare member.\n- **A module-private const** — never exported, so nothing resolves; backticks.\n- **A re-exported cross-package `Schema.Class`, `{@link}`ed from a file that\n  only `import type`s it** — even when the entry point re-exports the symbol\n  (`export { Package } from \"@effected\u002Fpackage-json\"`), a `{@link Package}`\n  written in a module that reaches it via `import type` fails with a\n  **different message** than the other cases: *\"This type of declaration is\n  not supported yet by the resolver\"*, and issues.json may attribute it to an\n  unrelated line through the rollup. Backticks are the fix. Confirmed at the\n  `@effected\u002Fsbom` build; the re-export itself is fine — only\n  the `{@link}` form is rejected.\n- **A symbol exported only from the root `.`, `{@link}`ed from a module that\n  ships under a second entry point** (`.\u002Ftesting`, `.\u002Fui`, `.\u002Fnode`) — each\n  entry point is its own API Extractor pass with its own export list (see \"A\n  second published entrypoint models its own surface\" below), so the root's\n  `{@link Foo}` is unresolved from `.\u002Ftesting`'s pass however public `Foo` is\n  at `.`. This is the most common way a multi-entrypoint package breaks the\n  rule. Backticks — or, when the symbol belongs to that surface, re-export it\n  from the second entry point too, and the link resolves there.\n\nThe diagnostic distinguishes the fixable case from the unfixable one, so read it\nrather than pattern-matching: *\"ambiguous… add a TSDoc member reference\nselector\"* means a selector will work (below), with one exception: a class\nmember whose name is declared both `static` and on the instance (case 4 under\n\"Backtick spans\" below); *\"does not have an export\"* means\nno selector ever will; *\"not supported yet by the resolver\"* means the resolver\ncannot follow the declaration shape at all — backticks, no selector to try.\n\n**Merged value + type names: use a member-reference selector, not a backtick.**\n\nAny name carrying **both a value and a type declaration** — an `interface` plus\na `const` of the same name (`ConfigCodec`, `MergeStrategy`, `VersionAccess`), or\na branded scalar's `const` schema plus its exported `type` (`PackageName`,\n`SpdxLicense`) — cannot be disambiguated by API Extractor from a bare link. Both\n`{@link X}` and `{@link X.member}` resolve to `ae-unresolved-link`, and the\ndiagnostic says so literally: *\"the reference is ambiguous… you need to add a\nTSDoc member reference selector.\"*\n\nThe fix is the selector, and the link keeps working:\n\n```ts\n\u002F** Wrap any {@link (ConfigCodec:interface)} with AES-GCM encryption. *\u002F\n\u002F** See {@link (ConfigResolver:variable).staticDir}. *\u002F\n\u002F** The decoded form of {@link (ConfigEventPayload:variable)}. *\u002F\n\u002F** A branded scalar: {@link (PackageName:type)}. *\u002F\n```\n\nPick the selector that names the declaration you mean: `:interface`, `:variable`,\n`:class`, `:type`. Getting it wrong still emits `ae-unresolved-link` — `:type` on\nan interface does not resolve — so a zero-warning build is the proof you chose\ncorrectly. `@effected\u002Fyaml` (`{@link (YamlSegment:type)}`) and\n`@effected\u002Fconfig-file` (`{@link (ConfigCodec:interface)}`) both ship these\nzero-warning.\n\n**Do not \"fix\" these by deleting the link.** Replacing a resolvable\n`{@link (X:interface)}` with an inert backtick span silently removes an API-doc\ncross-reference; a selector form resolves for every merged name.\n\n**Member of a const-only namespace object: link through the exported name — no\nselector, and the bare member name never resolves.** Every package entry point\nin this repo exports the shape\n\n```ts\nconst ascend = (path: string): string => path; \u002F\u002F module-local, NOT exported\nexport const Walker = { ascend } as const;     \u002F\u002F the only export\n```\n\nHere `{@link ascend}` is `ae-unresolved-link` — not because the name is\nambiguous, but because **there is no export by that name**, so no selector can\nfix it (`{@link (ascend:variable)}` stays red). Do not pattern-match \"bare link\nunresolved ⇒ reach for a selector\"; that heuristic belongs to the merged-name\ncase above, where the diagnostic says \"ambiguous\". When the diagnostic says\n*\"does not have an export\"*, the fix is a plain member reference through the\nexported const — `{@link Walker.ascend}` — with no selector, because `Walker`\ncarries a value declaration only. `packages\u002Fconfig-file\u002Fsrc\u002FConfigMigration.ts`\n(`{@link ConfigMigration.make}`) and `packages\u002Fwalker\u002Fsrc\u002FWalker.ts` both ship\nthis form warnings-clean. Expect the mistake to recur: `Jsonc`,\n`ConfigResolver`, `ConfigMigration`, `ConfigCodec`, `MergeStrategy`, and\n`Walker` all use this export shape, and the failure is invisible to `pnpm test`\nand `types:check` — it only surfaces in the **prod** build's `issues.json`.\n\n**Backtick spans remain correct for four cases**, where no selector helps:\n\n1. **Inherited members** — `{@link SemVer.make}` where `make` comes from the\n   synthesized base. There is no selector for a member the declaration does not\n   own. Write `` `SemVer.make` ``.\n2. **Cross-package symbols you only `import type`.** A selector disambiguates\n   *between local declarations*; it cannot reach a symbol API Extractor never\n   rolled up into this package entry point. An adapter that does\n   `import type { ConfigCodec } from \"@effected\u002Fconfig-file\"` and writes\n   `{@link (ConfigCodec:interface)}` still gets `ae-unresolved-link`. Write\n   `` `ConfigCodec` ``. Expect this on **every** adapter package that plugs\n   into a seam it does not own.\n3. **A `Schema.Class` FIELD** — `{@link ReleaseAgeGate.exclude}` or\n   `{@link WorkspaceCatalogs.set}`, where the member is a schema-declared field\n   rather than a hand-written class member. API Extractor does not see fields\n   declared in the factory's field object as class members at all, so **no\n   selector form fixes it** — `{@link (ReleaseAgeGate:class).exclude}` stays\n   `ae-unresolved-link` too. Write `` `ReleaseAgeGate.exclude` ``. Note the asymmetry with the\n   inherited-member case above — there the *declaration* does not own the\n   member; here the class does own it, but only through the factory, which is\n   the same dead end from API Extractor's side.\n4. **A class with a same-named `static` and instance member** — a\n   `Fn.dual` static `Range.filter` beside an instance `filter(versions)`.\n   `{@link Range.filter}` reports *\"ambiguous because \"filter\" has more than\n   one declaration; you need to add a TSDoc member reference selector\"*, and\n   then rejects every selector you can write: `{@link Range.(filter:static)}`\n   and `{@link Range.(filter:instance)}` are both *Unsupported system\n   selector*. The advice in the diagnostic is a dead end. The house answer is\n   to **not create the collision**: keep the instance method canonical and\n   drop the static (derive a data-last form from an `Order` or a lambda), or\n   give the static a distinct name. `@effected\u002Fsemver` dropped its seven\n   comparison statics and `Range.filter` for this reason. Where a collision\n   cannot be avoided, write the name in backticks.\n\n## Overloaded exported functions: the index selector, parenthesized\n\nAdding a second overload to an exported function is an ordinary additive\nchange, and it breaks every existing `{@link fn}` at once with the same\n*\"ambiguous… add a TSDoc member reference selector\"* diagnostic the merged-name\ncase emits. The selector that fixes it is **not** `:function` — it is the\n**overload index**, 1-based in `.d.ts` declaration order, and the whole\nreference must be parenthesized:\n\n```ts\n\u002F** The recipe form of {@link (descend:1)}; under `\"record\"` see {@link (descend:2)}. *\u002F\n```\n\nProbed at the `@effected\u002Fwalker` prod build (`descend` carries two\noverload signatures plus an implementation signature), every form in one\ncomment so the controls and the answer share a run:\n\n| Form | Result |\n| --- | --- |\n| `{@link (descend:1)}`, `{@link (descend:2)}` | resolve, no warning |\n| `{@link descend}` | *ambiguous… add a TSDoc member reference selector* |\n| `{@link (descend:function)}` | *More than one declaration \"descend\" matches the TSDoc selector \"function\"* |\n| `{@link descend:1}` | *Syntax error… the member selector must be enclosed in parentheses* |\n| `{@link (descend:3)}` | *An overload for \"descend\" was not found that matches the TSDoc selector \":3\"* |\n\nThree consequences worth holding onto:\n\n- **The system selector cannot disambiguate overloads.** Every overload is a\n  `FunctionDeclaration`, so `:function` matches all of them and fails on the\n  count — reach for the number, not the kind.\n- **The implementation signature does not count.** `:3` fails on a function\n  with two overloads and an implementation, because the index walks the\n  rolled-up `.d.ts`, where only the overload signatures survive\n  (`AstReferenceResolver.#selectUsingIndexSelector` →\n  `Collector.getOverloadIndex`, which numbers same-kind declarations of the\n  symbol in order, from 1).\n- **Reordering overloads renumbers links.** `:1` names *position*, not\n  signature, so a refactor that swaps two overloads silently repoints every\n  link without a warning — the build stays green either way. Prefer citing the\n  overload whose position is stable, and re-read the links when you reorder.\n\nBackticks still work here. Where the link would carry the reader to the right overload, use the\nselector; where any overload will do and the position is likely to move, a\nbacktick is the honest form.\n\n## `{@link X.member}` where the member name IS a TSDoc selector keyword\n\nA member whose name collides with a TSDoc **system selector** (`variable`,\n`function`, `class`, `interface`, `namespace`, `enum`, `constructor`,\n`static`, `instance`, …) makes `{@link ActionInput.variable}` ambiguous to\nthe parser, and the bundler warns *\"identifier must be quoted\"*. Hit live on a\nstatic genuinely named `variable`. The fix is the same as\nevery other unresolvable-link case in this skill: **backticks**\n(`` `ActionInput.variable` ``) — do not try quoted-selector syntax, and do\nnot rename the API to appease the doc engine.\n\n## `tsdoc-at-sign-in-word`: never markdown-link a scoped package\n\nA scoped package name in TSDoc must be backticked. That much is known — the trap\nis that **a markdown link hides a second, unbackticked copy inside its URL**, and\nthat copy is the one that trips:\n\n```ts\n\u002F** See [@effected\u002Fconfig-file](https:\u002F\u002Fnpmjs.com\u002Fpackage\u002F@effected\u002Fconfig-file). *\u002F\n\u002F\u002F     ^ backticked? no — but even fixing the label leaves the URL's copy\n```\n\nThe label being wrapped in `[...]` is not backticking, and the `@effected` inside\nthe parenthesized URL is bare text as far as the TSDoc parser is concerned.\nWorse, the diagnostic **reports at the NEXT declaration's line**, so the reported\nposition points at an innocent symbol below the comment — the same off-by-a-\ndeclaration behaviour as the `issues.json` note below, compounded.\n\n**The rule: backticks only. Never markdown-link a scoped package name in TSDoc**\n— not in the label, not in the URL. If the URL genuinely matters, put the bare\nURL on its own (`{@link https:\u002F\u002F…}` or plain text) and name the package in a\nbacktick span beside it.\n\n**A note on reading `issues.json`.** Its `file` names the source, but its `line`\nindexes the **generated `.d.ts`**. Locate a `tsdoc-*` or `ae-unresolved-link`\ndefect textually, not by jumping to that line number — the position routinely\nlands on a class declaration thirty lines from the offending comment, which has\nsent more than one agent chasing an innocent symbol.\n\n## `tsdoc-code-span-missing-delimiter`: a code span must not wrap\n\nA backtick code span must open and close on the **same comment line**. Prose\nreflow is how this breaks: a span like\n`` `Unsupported package manager specification` `` wrapped across two comment\nlines reads perfectly balanced in source — each line's fragment looks fine —\nbut the parser sees two half-spans and emits **two**\n`tsdoc-code-span-missing-delimiter` warnings in the prod gate. When wrapping\nwould split a span, rewrap the sentence, not the span: move the whole span to\nthe next line, or shorten the prose around it. (Found live in\n`@effected\u002Fpackage-json`, where an error-message quotation\nwrapped during an ordinary TSDoc edit.)\n\nThe warning **cascades and misdirects**: once a span is unterminated, the\nbraces inside it go raw, so a single wrapped span like\n`` `PATCH \u002Frepos\u002F{owner}\u002F{repo}` `` fans out into `tsdoc-escape-right-brace`\nand `tsdoc-malformed-inline-tag` warnings whose reported lines sit BELOW the\ncomment (declaration-relative reporting). Do not chase those by escaping\nbraces — a properly closed one-line span protects `{`\u002F`}` fine, and\nbrace-bearing spans like `` `{ status: \"enabled\" }` `` are legal as-is. Find\nthe one wrapped span and rejoin it; the whole fan collapses. (Bitten twice — 14 warnings in `@effected\u002Fgit`, 5 in `@effected\u002Fgithub`, each\none wrapped argv\u002Froute literal; a reader triaging the second incident from\nthe brace warnings alone diagnosed brace-escaping, which is exactly the wrong\nfix.)\n\n## A bin-only package has no API model — `emitDts: false`, not a fake entry point\n\nA package whose published surface is a `bin` and nothing else (`exports`\nlimited to `\".\u002Fpackage.json\"`, no `index.ts`) fails the build under the\ndefault config with `Cannot merge zero API models`. Nothing is wrong\nwith the package: the declaration pass ran over zero entry points and the\nprod meta pass then had nothing to extract. The fix is **not** an `index.ts`\nthat exports nothing, and not a `meta: false` (which still runs the dts\npass). It is `emitDts = false` in the package's `effected.toml`:\n\n```toml\n# Bin-only: no declarations to bundle, no API model to extract.\n[package.build]\nemitDts = false\n```\n\n`BuildConfigInput.emitDts` is documented as the switch for *\"JS-only artifacts\nthat never consume declarations (e2e fixtures, bins, internal tools)\"* — it\nskips the dts pass (dev and prod) and therefore the meta pass, while still\nemitting JS, the byte-variant targets and the transformed `package.json`.\nSuch a package therefore carries **no `_base` suppression, no `tsdoc.json`, no\napi-extractor model and no website page** — all of that machinery reads a\n`.d.ts` that does not exist. It is also the one package shape a\ndoc-model-driven enumerator (a construct index, a docs site) must exclude by\nits `exports` map (every key `\".\u002Fpackage.json\"`), or it reports the bin as a\nperpetually missing build. The `effect-v4-cli` skill covers the package shape\nitself.\n\n## A second published entrypoint models its own surface\n\nA `package.json` `exports` map can publish more than one subpath — `.` and\n`.\u002Ftesting`, say — and api-extractor builds and gates **each entrypoint\nindependently**, as its own closed surface. A type that one entrypoint's\npublic signature reaches is not automatically visible through another\nentrypoint just because both ship from the same package: every kit type an\nentrypoint's own exported signatures name must be re-exported from **that**\nentrypoint too, or that entrypoint alone reports `ae-forgotten-export` —\n\"already exported from `.`\" is never the fix for a second entrypoint's own\ngate.\n\nWorked example: `@effected\u002Fworkspaces\u002Ftesting`\n(`packages\u002Fworkspaces\u002Fsrc\u002Ftesting.ts`). `WorkspaceLayering.checkWorkspace`'s\nrequirement and error channel reach the `WorkspaceDiscovery` service — its\nshape, options and failures. `WorkspaceDiscovery` is not part of the\npackage's main `.` surface at all; `.\u002Ftesting` re-exports it anyway, because\n`.\u002Ftesting`'s own signatures are the ones naming it. The module's own header\ncomment states the rule: \"This module is an ENTRY POINT: api-extractor\nmodels it as its own surface, so every kit type its signatures name is\nre-exported here.\"\n\n```ts\nimport type { WorkspaceDiscovery, WorkspaceLayering } from \"@effected\u002Fworkspaces\u002Ftesting\"\n\n\u002F\u002F Both resolve from the SAME subpath: WorkspaceLayering's own signatures\n\u002F\u002F reach WorkspaceDiscovery (edgesOf's parameter, checkWorkspace's\n\u002F\u002F requirement and error channel), so this entrypoint re-exports\n\u002F\u002F WorkspaceDiscovery too rather than relying on the package's main entry\n\u002F\u002F to already have done so.\ndeclare const service: WorkspaceDiscovery\ndeclare const layering: typeof WorkspaceLayering\n```\n\nA second entrypoint that re-exports types whose TSDoc `{@link}`s a class only\nthe main entry exports fails `ae-unresolved-link` on the second entry's own\nreport: the link has nothing to resolve against there. Spell those references\nin backticks instead of `{@link}`. `packages\u002Fcommands\u002FCLAUDE.md` and\n`packages\u002Fmemfs` both do this.\n\n### A subpath that names the root's types: the self-import pattern\n\nA second entrypoint whose signatures name types the root already exports —\n`@effected\u002Fnpm\u002Fnode` names `NpmConfig`, `@effected\u002Fpnpm\u002Fsettings` names\n`PnpmSettings`, `PeerDependencyRules` and `VersioningConfig` — should import\nthem from the package's own name rather than carry a copy. Three parts, all\nrequired:\n\n1. **Keep the self-import external in the declarations**, in the package's\n   `effected.toml`: `[package.build] dtsExternals = [\"@effected\u002Fnpm\"]`. The\n   subpath's `.d.ts` then refers to the root's type instead of inlining it.\n2. **Scope a suppression to that entry point only.** The second API Extractor\n   pass resolves the self-import to the built root declarations but still\n   reports the root types it reaches as forgotten, so add\n   `{ messageId = \"ae-forgotten-export\", pattern = 'entry point node\\.d\\.ts$' }`\n   (the subpath's own file name) to `[package.build.meta.tsdoc]\n   suppressWarnings`. A bare pattern would hide a real leak in the root.\n3. **Pin what the suppression could hide with a built-declaration test.** The\n   suppression is never unpinned: a declaration local to the subpath and left\n   unexported would be silenced, so a test reads the built `\u003Centry>.d.ts` and\n   asserts that every top-level declaration is exported and the import\n   specifiers are the expected ones (`__test__\u002Fnode-declarations.test.ts` in\n   `@effected\u002Fnpm`, `__test__\u002Fsettings\u002Fdeclarations.test.ts` in\n   `@effected\u002Fpnpm`). Build before the test reads it.\n\n## Reading the gate without fooling yourself\n\n`issues.json` is a **false-green oracle**. Four rules, each learned by being burned:\n\n1. **Build through Turbo.** `pnpm build --filter \u003Cpkg>` from the repo root runs\n   `types:check` first and caches. A direct `effected package build` in the\n   package directory is allowed but skips that gate and the cache, so a type\n   error the typecheck would have caught surfaces as an API Extractor failure\n   instead.\n\n2. **A crashed build can leave every bucket empty** — byte-shaped exactly like\n   a perfectly clean gate. **Read `buildOk` first**: `false` (with `failure`\n   saying why) or a missing stamp means the buckets are not evidence. An\n   unexpected `suppressed` count is a second tell: counting only\n   `source: \"api-extractor\"` entries (tsdown advisory noise such as\n   `TSGO_EXPERIMENTAL` lands there too), a package with class factories has at\n   least one `_base` entry per factory. Never conclude \"clean\"\n   from empty buckets alone; `building-a-package`'s reading-issues reference\n   has the full read order.\n\n3. **An incremental build can hide warnings.** A stale `dist\u002F.tsbuildinfo.lib`\n   feeds API Extractor an old `.d.ts`, so a warning present in a cold build is\n   absent from the incremental one. Read the gate from a cold build (`rm -rf`\n   the package's `dist` first) whenever the answer matters.\n\n4. **Concurrent agents corrupt each other's gates.** `dist\u002F` has no lock:\n   two sessions building the same package (or one running `rm -rf dist &&\n   build` while another reads) produce torn artifacts that read as rule 2's\n   clean-shaped truncation — what looks like a nondeterministic bundler flake\n   is usually contention. Treat any `issues.json` read as suspect unless it\n   immediately follows YOUR serial, cold, exit-0 build of that package. And\n   **never `pkill -f` a shared build-tool pattern** (`pkill -f \"effected package build\"`\n   killed other sessions' legitimate builds — the actual root cause of one\n   \"flake\" storm); kill your own child process by PID or not at all.\n\n`dist\u002F` is gitignored and shared. It carries no commit, no cleanliness, and no\nexit-code provenance, so it is only evidence when *your* build exited 0 against\na clean tree with nothing else building. If you mutate source to prove a fix is\nload-bearing, **rebuild afterward** — otherwise you leave an artifact that\ndescribes a tree no commit ever contained.\n\nOne cosmetic non-defect while reading built declarations: core's dollar-alias\nexport names surface verbatim — a `Schema.Array` field emits as\n`Schema.$Array\u003C…>`, `Schema.Record` as `Schema.$Record\u003C…>`. Correct output,\nobserved across multiple package builds; do not \"fix\" it, and do\nnot read it as a wrong import in the source.\n\n## Related skills\n\n`authoring-tsdoc` owns the general policy this skill assumes: the binary\nrelease-tag rule, the supported-tag allow-list, custom tags, when a\nsuppression is legitimate, the `typeof local` forgotten export, and every\nother `ae-*`\u002F`tsdoc-*` fix. `building-a-package` owns `effected.toml` and the\nbuild itself. The `tsdoc-fixer` agent carries all three and runs the fix loop\nend to end.\n\n## Rejected alternatives\n\nTwo other idioms are not used: the `@internal`-tagged base (it leaves a\nresidual non-fatal `ae-incompatible-release-tags`) and the `@public X_base`\nconst-with-explicit-annotation idiom (the inline form + scoped `_base`\nsuppression covers the recursive and `Context.Service` cases it existed for).\n\nThe inline factory + scoped `_base` suppression is the **single policy**, and\nthe transitional backlog is **drained**: no package in the kit carries a\n`@public X_base` const. `grep -rln \"_base\" packages\u002F*\u002Fsrc\u002F` returns a single\nhit — a *comment* in `runtimes\u002Fsrc\u002FNodeResolver.ts` explaining why the\nsuppression is scoped. The suppression is the build's default, so no package needs to carry the line unless it overrides\n`meta.tsdoc.suppressWarnings`. Find the packages that override it live:\n\n```bash\ngrep -l \"suppressWarnings\" packages\u002F*\u002Feffected.toml\n```\n\nEvery package that grep prints must still list the `_base` rule. If you find a\n`@public X_base` const anywhere, it is new code written against the retired\nidiom, not residue: inline the factory into the heritage clause and delete\nthe const.\n","# API Extractor × Effect class factories\n\n`class X extends Schema.Class\u003CX>(\"X\")({...}) {}` (and `TaggedClass` \u002F\n`TaggedError` \u002F `Opaque` \u002F `Context.Service`) produces an anonymous\nheritage type. API Extractor reports it as `ae-forgotten-export` on a\nsynthesized `X_base` symbol — CI-fatal under `effected package build`.\n\n## The policy: inline factory + scoped `_base` suppression\n\nWrite the factory **inline** — no split-out base const, no hand-written\nannotation — and suppress the synthesized-base warning narrowly. The build's\ndefault `meta.tsdoc.suppressWarnings` is exactly that one rule, so a package\nwith no `effected.toml` already has it. A package that sets its own list\nreplaces the default and must restate the rule in its `effected.toml`:\n\n```toml\n[package.build.meta.tsdoc]\nsuppressWarnings = [{ messageId = \"ae-forgotten-export\", pattern = \"_base\" }]\n```\n\n```text\n\u002F\u002F src\u002FX.ts — inline, no base const, no annotation\n\u002F** ... @public *\u002F\nexport class SemVer extends Schema.Class\u003CSemVer>(\"SemVer\")({\n  major: nonNegativeInteger,\n  minor: nonNegativeInteger,\n  patch: nonNegativeInteger,\n}) {\n  \u002F** ... *\u002F\n  static readonly FromString = \u002F* ... *\u002F;\n}\n```\n\nThat yields `dist\u002Fprod\u002Fissues.json` = `{ warnings: 0, errors: 0, suppressed: N }`,\nwith the `SemVer_base` (and every other `*_base`) entry in the **`suppressed`**\nbucket.\n\n### Why this is principled, not a \"silence the warning\" cop-out\n\nThe emitted `.d.ts` proves the base is genuinely internal:\n\n```text\ndeclare const SemVer_base: Schema.Class\u003CSemVer, \u002F* full inlined shape *\u002F>;\ndeclare class SemVer extends SemVer_base { \u002F* full shape: fields, statics, methods *\u002F }\nexport { SemVer \u002F* , ... *\u002F };   \u002F\u002F \u003C- SemVer_base is NOT exported\n```\n\n`X_base` is a module-local `declare const` that never appears in the export\nlist, and the exported class carries its full shape. A consumer can neither\nname nor need `X_base`. The suppression is **narrow** (`pattern: \"_base\"`\nmatches only these synthesized symbols), **accounted for** (it lands in the\n`suppressed` bucket and the build log reports the count — it is logged, not\nsilenced), and **harmless** (`ae-forgotten-export` means \"a public API names a\ntype a consumer can't import\"; here nothing a consumer needs is lost). That is\ncategorically different from muting, say, an `ae-missing-release-tag` you don't\nfeel like fixing.\n\n### Validated across every factory kind\n\nConfirmed clean (`warnings: 0`, the `*_base` in `suppressed`) with `tsgo`\ngreen and **no hand-written annotations** on:\n\n- `Schema.Opaque`, `Schema.Class`, `TaggedClass`, `TaggedError`.\n  (Naming trap: there is no `Schema.asClass` — zero occurrences in\n  `Schema.ts`. To give a schema value a class identity you subclass it\n  directly: `class MyString extends Schema.String {}`.)\n- **Recursive `Schema.Class` + `Schema.suspend`** (a node whose field\n  references itself): the inline form has **no TS2506** and needs none of the\n  old `Schema.Schema\u003CSelf>` base-annotation gymnastics — that failure mode was\n  an artifact of splitting the annotated base const out of the heritage clause,\n  and it evaporates inline. The suspend callback still takes its normal\n  `(): Schema.Schema\u003CSelf> => Self` return type; that is ordinary recursive-schema\n  practice, not a base annotation.\n- **`Context.Service`** (`class R extends Context.Service\u003CR, Shape>()(\"id\") {}`):\n  same `R_base` suppression, no annotation.\n- **Named field-schema consts** (`major: nonNegativeInteger`): the const's type\n  **inlines into `X_base`** rather than emitting as `typeof nonNegativeInteger`,\n  so it does **not** forgotten-export and does **not** need `@public`. The old\n  \"schema helpers referenced by the annotation become `@public`\" cascade retires\n  with the annotation that caused it.\n\nSo the former ceremony — the `@public X_base` export, the mandatory explicit\nannotation, the recursive `Schema.Schema\u003CSelf>` special-casing, re-exporting\nevery base from `index.ts`, and marking field consts `@public` — is all gone.\nWrite the class the way Effect's own docs write it and add the one-line\nsuppression.\n\n## The suppression is `_base`-scoped ONLY — it does not cover method signatures\n\nThe binary release-tag rule still applies to **method and function\nparameter\u002Freturn types**: *anything a `@public` signature names must itself be\n`@public`*. An internal type on a public method signature is a **different\nsymbol** (not suffixed `_base`), so `pattern: \"_base\"` does **not** suppress it —\nby design, so genuine surface leaks are never masked. In the yaml port, one\ninternal `RawDiagnostic` parameter on a `@public`\n`YamlDiagnostic.fromRaw(raw: RawDiagnostic, …)` produced **12**\nforgotten-export warnings (the whole `internal\u002Fdiagnostics` module). These are\nreal and must be fixed, not suppressed:\n\n- **Inline a structural type** on the public signature so no internal symbol is\n  named — best for engine-internal record types that should not become public\n  surface:\n\n  ```text\n  static fromRaw(\n   raw: { readonly code: YamlErrorCode; readonly message: string; readonly offset: number; readonly length: number },\n   text: string,\n  ): YamlDiagnostic { … }\n  ```\n\n  (This is the **landed** form, not a hypothetical: `@effected\u002Fyaml`'s\n  `YamlDiagnostic.fromRaw` ships exactly this structural-inline signature — the\n  12-warning `RawDiagnostic` incident above is what motivated it.)\n\n- **Or tag the referenced type `@public`** and re-export it — only when it is\n  genuinely part of the API.\n\nPrefer the structural-inline form for anything under `src\u002Finternal\u002F`. Watch for\nthis on `X.fromRaw` \u002F `X.of` \u002F codec-adapter statics that bridge the internal\nengine to the public classes — exactly where an internal record type sneaks\nonto a `@public` signature.\n\n**Three escapes that look like they should work and do not.** Two build cycles\nwere burned re-discovering this, so it is worth stating outright:\n\n| Attempted escape | Result |\n| --- | --- |\n| Mark the *method* `@internal` and leave it on the `@public` class | **Still fails.** The gate follows the class's release tag; an `@internal` method of a `@public` class is still reachable rollup surface. |\n| Mark the *referenced type* `@internal` | **Still fails**, and differently — now you get `ae-incompatible-release-tags` instead. `@internal` never resolves an `ae-forgotten-export`; it only relabels the problem. |\n| Extract the shape to a named `type` alias and use that | **Still fails.** An alias is a named symbol like any other, so it forgotten-exports under its own name. Naming the thing is precisely what you cannot do. |\n\nOnly two things actually work: **inline the type structurally** (no name exists,\nso nothing can be forgotten), or **export it** and accept it as public API. There\nis no third option, and the effort spent looking for one is unrecoverable.\n\n### The positive counterpart: third-party generics are free\n\nThe mirror-image fear — that naming a *dependency's* generic type on a `@public`\nsignature will forgotten-export it — is unfounded. **API Extractor resolves\ndeclared-dependency third-party types as externals and emits real `import`\nstatements in the `.d.ts`**: no `ae-forgotten-export`, no inlining, no\nre-export ceremony. Confirmed at `@effected\u002Fgithub`'s first build with octokit's\n`Endpoints[R]` on a `@public` signature.\n\nThe condition is that the package be a **declared dependency** (it must be\nresolvable from the emitted `.d.ts`, which a `devDependency` or a transitive\npackage is not). Given that, write the precise third-party type rather than\nwidening to `unknown` or hand-copying its shape — the gate does not charge you\nfor it.\n\n### …nor does it cover a type in the class factory's GENERIC\n\nThe same gap bites one step earlier, in the **heritage clause itself**. An\nunexported interface passed as a *type argument* to the factory —\n\n```ts\nimport { Context } from \"effect\"\n\ninterface VersionCacheShape {\n  readonly current: string\n} \u002F\u002F NOT exported from index.ts\n\nexport class VersionCache extends Context.Service\u003CVersionCache, VersionCacheShape>()(\n \"@effected\u002Fsemver\u002FVersionCache\",\n) {}\n```\n\n— leaks as a plain `ae-forgotten-export` on **`VersionCacheShape`**, whose symbol\nname has no `_base` suffix, so `pattern: \"_base\"` does not match it.\n\n**Measured** on `@effected\u002Fsemver` by un-exporting exactly that interface and\nrebuilding:\n\n| | warnings | suppressed | code | `ciFatal` |\n| --- | --- | --- | --- | --- |\n| baseline | 0 | 12 | — | — |\n| interface un-exported | **1** | 12 (unchanged) | `ae-forgotten-export` | **`true`** |\n\n> `The symbol \"VersionCacheShape\" needs to be exported by the entry point index.d.ts`\n\nThe suppressed bucket stayed at 12 and every entry in it ends in `_base` — the\nnarrow suppression does exactly what it says and nothing more.\n\nThe fix is to **export the shape type** from the entrypoint (it is genuinely part\nof the API — it is the service's contract), not to widen the suppression. This is\nthe case a reviewer walks you into: nitting \"dedupe this shape into a named\ninterface\" is right on the merits, but the moment the name exists it must also be\nexported, or the gate goes red. Applies equally to `Schema.Class\u003CSelf, Fields>`\nand every other class factory that takes a type argument.\n\n## TSDoc `{@link}` traps\n\n**The one rule the cases below are instances of:**\n\n> `{@link}` resolves **only to symbols the entry point exports, under the name it\n> exports them by**. Everything else is a backtick span.\n\nSix ways a name fails that test:\n\n- **A schema-declared `Schema.Class` field** — the class owns it, but only\n  through the factory, which API Extractor does not see through.\n- **A module-local renamed at export** — `class Route` exported as\n  `export { Route as RestRoute }` resolves as `{@link RestRoute}` and **never**\n  as `{@link Route}`, no matter that `Route` is the name in the file you are\n  writing the comment in. Confirmed twice at the `@effected\u002Fgithub` build.\n- **A member of a const-only namespace object** — link through the exported\n  const (`{@link Walker.ascend}`), never the bare member.\n- **A module-private const** — never exported, so nothing resolves; backticks.\n- **A re-exported cross-package `Schema.Class`, `{@link}`ed from a file that\n  only `import type`s it** — even when the entry point re-exports the symbol\n  (`export { Package } from \"@effected\u002Fpackage-json\"`), a `{@link Package}`\n  written in a module that reaches it via `import type` fails with a\n  **different message** than the other cases: *\"This type of declaration is\n  not supported yet by the resolver\"*, and issues.json may attribute it to an\n  unrelated line through the rollup. Backticks are the fix. Confirmed at the\n  `@effected\u002Fsbom` build; the re-export itself is fine — only\n  the `{@link}` form is rejected.\n- **A symbol exported only from the root `.`, `{@link}`ed from a module that\n  ships under a second entry point** (`.\u002Ftesting`, `.\u002Fui`, `.\u002Fnode`) — each\n  entry point is its own API Extractor pass with its own export list (see \"A\n  second published entrypoint models its own surface\" below), so the root's\n  `{@link Foo}` is unresolved from `.\u002Ftesting`'s pass however public `Foo` is\n  at `.`. This is the most common way a multi-entrypoint package breaks the\n  rule. Backticks — or, when the symbol belongs to that surface, re-export it\n  from the second entry point too, and the link resolves there.\n\nThe diagnostic distinguishes the fixable case from the unfixable one, so read it\nrather than pattern-matching: *\"ambiguous… add a TSDoc member reference\nselector\"* means a selector will work (below), with one exception: a class\nmember whose name is declared both `static` and on the instance (case 4 under\n\"Backtick spans\" below); *\"does not have an export\"* means\nno selector ever will; *\"not supported yet by the resolver\"* means the resolver\ncannot follow the declaration shape at all — backticks, no selector to try.\n\n**Merged value + type names: use a member-reference selector, not a backtick.**\n\nAny name carrying **both a value and a type declaration** — an `interface` plus\na `const` of the same name (`ConfigCodec`, `MergeStrategy`, `VersionAccess`), or\na branded scalar's `const` schema plus its exported `type` (`PackageName`,\n`SpdxLicense`) — cannot be disambiguated by API Extractor from a bare link. Both\n`{@link X}` and `{@link X.member}` resolve to `ae-unresolved-link`, and the\ndiagnostic says so literally: *\"the reference is ambiguous… you need to add a\nTSDoc member reference selector.\"*\n\nThe fix is the selector, and the link keeps working:\n\n```ts\n\u002F** Wrap any {@link (ConfigCodec:interface)} with AES-GCM encryption. *\u002F\n\u002F** See {@link (ConfigResolver:variable).staticDir}. *\u002F\n\u002F** The decoded form of {@link (ConfigEventPayload:variable)}. *\u002F\n\u002F** A branded scalar: {@link (PackageName:type)}. *\u002F\n```\n\nPick the selector that names the declaration you mean: `:interface`, `:variable`,\n`:class`, `:type`. Getting it wrong still emits `ae-unresolved-link` — `:type` on\nan interface does not resolve — so a zero-warning build is the proof you chose\ncorrectly. `@effected\u002Fyaml` (`{@link (YamlSegment:type)}`) and\n`@effected\u002Fconfig-file` (`{@link (ConfigCodec:interface)}`) both ship these\nzero-warning.\n\n**Do not \"fix\" these by deleting the link.** Replacing a resolvable\n`{@link (X:interface)}` with an inert backtick span silently removes an API-doc\ncross-reference; a selector form resolves for every merged name.\n\n**Member of a const-only namespace object: link through the exported name — no\nselector, and the bare member name never resolves.** Every package entry point\nin this repo exports the shape\n\n```ts\nconst ascend = (path: string): string => path; \u002F\u002F module-local, NOT exported\nexport const Walker = { ascend } as const;     \u002F\u002F the only export\n```\n\nHere `{@link ascend}` is `ae-unresolved-link` — not because the name is\nambiguous, but because **there is no export by that name**, so no selector can\nfix it (`{@link (ascend:variable)}` stays red). Do not pattern-match \"bare link\nunresolved ⇒ reach for a selector\"; that heuristic belongs to the merged-name\ncase above, where the diagnostic says \"ambiguous\". When the diagnostic says\n*\"does not have an export\"*, the fix is a plain member reference through the\nexported const — `{@link Walker.ascend}` — with no selector, because `Walker`\ncarries a value declaration only. `packages\u002Fconfig-file\u002Fsrc\u002FConfigMigration.ts`\n(`{@link ConfigMigration.make}`) and `packages\u002Fwalker\u002Fsrc\u002FWalker.ts` both ship\nthis form warnings-clean. Expect the mistake to recur: `Jsonc`,\n`ConfigResolver`, `ConfigMigration`, `ConfigCodec`, `MergeStrategy`, and\n`Walker` all use this export shape, and the failure is invisible to `pnpm test`\nand `types:check` — it only surfaces in the **prod** build's `issues.json`.\n\n**Backtick spans remain correct for four cases**, where no selector helps:\n\n1. **Inherited members** — `{@link SemVer.make}` where `make` comes from the\n   synthesized base. There is no selector for a member the declaration does not\n   own. Write `` `SemVer.make` ``.\n2. **Cross-package symbols you only `import type`.** A selector disambiguates\n   *between local declarations*; it cannot reach a symbol API Extractor never\n   rolled up into this package entry point. An adapter that does\n   `import type { ConfigCodec } from \"@effected\u002Fconfig-file\"` and writes\n   `{@link (ConfigCodec:interface)}` still gets `ae-unresolved-link`. Write\n   `` `ConfigCodec` ``. Expect this on **every** adapter package that plugs\n   into a seam it does not own.\n3. **A `Schema.Class` FIELD** — `{@link ReleaseAgeGate.exclude}` or\n   `{@link WorkspaceCatalogs.set}`, where the member is a schema-declared field\n   rather than a hand-written class member. API Extractor does not see fields\n   declared in the factory's field object as class members at all, so **no\n   selector form fixes it** — `{@link (ReleaseAgeGate:class).exclude}` stays\n   `ae-unresolved-link` too. Write `` `ReleaseAgeGate.exclude` ``. Note the asymmetry with the\n   inherited-member case above — there the *declaration* does not own the\n   member; here the class does own it, but only through the factory, which is\n   the same dead end from API Extractor's side.\n4. **A class with a same-named `static` and instance member** — a\n   `Fn.dual` static `Range.filter` beside an instance `filter(versions)`.\n   `{@link Range.filter}` reports *\"ambiguous because \"filter\" has more than\n   one declaration; you need to add a TSDoc member reference selector\"*, and\n   then rejects every selector you can write: `{@link Range.(filter:static)}`\n   and `{@link Range.(filter:instance)}` are both *Unsupported system\n   selector*. The advice in the diagnostic is a dead end. The house answer is\n   to **not create the collision**: keep the instance method canonical and\n   drop the static (derive a data-last form from an `Order` or a lambda), or\n   give the static a distinct name. `@effected\u002Fsemver` dropped its seven\n   comparison statics and `Range.filter` for this reason. Where a collision\n   cannot be avoided, write the name in backticks.\n\n## Overloaded exported functions: the index selector, parenthesized\n\nAdding a second overload to an exported function is an ordinary additive\nchange, and it breaks every existing `{@link fn}` at once with the same\n*\"ambiguous… add a TSDoc member reference selector\"* diagnostic the merged-name\ncase emits. The selector that fixes it is **not** `:function` — it is the\n**overload index**, 1-based in `.d.ts` declaration order, and the whole\nreference must be parenthesized:\n\n```ts\n\u002F** The recipe form of {@link (descend:1)}; under `\"record\"` see {@link (descend:2)}. *\u002F\n```\n\nProbed at the `@effected\u002Fwalker` prod build (`descend` carries two\noverload signatures plus an implementation signature), every form in one\ncomment so the controls and the answer share a run:\n\n| Form | Result |\n| --- | --- |\n| `{@link (descend:1)}`, `{@link (descend:2)}` | resolve, no warning |\n| `{@link descend}` | *ambiguous… add a TSDoc member reference selector* |\n| `{@link (descend:function)}` | *More than one declaration \"descend\" matches the TSDoc selector \"function\"* |\n| `{@link descend:1}` | *Syntax error… the member selector must be enclosed in parentheses* |\n| `{@link (descend:3)}` | *An overload for \"descend\" was not found that matches the TSDoc selector \":3\"* |\n\nThree consequences worth holding onto:\n\n- **The system selector cannot disambiguate overloads.** Every overload is a\n  `FunctionDeclaration`, so `:function` matches all of them and fails on the\n  count — reach for the number, not the kind.\n- **The implementation signature does not count.** `:3` fails on a function\n  with two overloads and an implementation, because the index walks the\n  rolled-up `.d.ts`, where only the overload signatures survive\n  (`AstReferenceResolver.#selectUsingIndexSelector` →\n  `Collector.getOverloadIndex`, which numbers same-kind declarations of the\n  symbol in order, from 1).\n- **Reordering overloads renumbers links.** `:1` names *position*, not\n  signature, so a refactor that swaps two overloads silently repoints every\n  link without a warning — the build stays green either way. Prefer citing the\n  overload whose position is stable, and re-read the links when you reorder.\n\nBackticks still work here. Where the link would carry the reader to the right overload, use the\nselector; where any overload will do and the position is likely to move, a\nbacktick is the honest form.\n\n## `{@link X.member}` where the member name IS a TSDoc selector keyword\n\nA member whose name collides with a TSDoc **system selector** (`variable`,\n`function`, `class`, `interface`, `namespace`, `enum`, `constructor`,\n`static`, `instance`, …) makes `{@link ActionInput.variable}` ambiguous to\nthe parser, and the bundler warns *\"identifier must be quoted\"*. Hit live on a\nstatic genuinely named `variable`. The fix is the same as\nevery other unresolvable-link case in this skill: **backticks**\n(`` `ActionInput.variable` ``) — do not try quoted-selector syntax, and do\nnot rename the API to appease the doc engine.\n\n## `tsdoc-at-sign-in-word`: never markdown-link a scoped package\n\nA scoped package name in TSDoc must be backticked. That much is known — the trap\nis that **a markdown link hides a second, unbackticked copy inside its URL**, and\nthat copy is the one that trips:\n\n```ts\n\u002F** See [@effected\u002Fconfig-file](https:\u002F\u002Fnpmjs.com\u002Fpackage\u002F@effected\u002Fconfig-file). *\u002F\n\u002F\u002F     ^ backticked? no — but even fixing the label leaves the URL's copy\n```\n\nThe label being wrapped in `[...]` is not backticking, and the `@effected` inside\nthe parenthesized URL is bare text as far as the TSDoc parser is concerned.\nWorse, the diagnostic **reports at the NEXT declaration's line**, so the reported\nposition points at an innocent symbol below the comment — the same off-by-a-\ndeclaration behaviour as the `issues.json` note below, compounded.\n\n**The rule: backticks only. Never markdown-link a scoped package name in TSDoc**\n— not in the label, not in the URL. If the URL genuinely matters, put the bare\nURL on its own (`{@link https:\u002F\u002F…}` or plain text) and name the package in a\nbacktick span beside it.\n\n**A note on reading `issues.json`.** Its `file` names the source, but its `line`\nindexes the **generated `.d.ts`**. Locate a `tsdoc-*` or `ae-unresolved-link`\ndefect textually, not by jumping to that line number — the position routinely\nlands on a class declaration thirty lines from the offending comment, which has\nsent more than one agent chasing an innocent symbol.\n\n## `tsdoc-code-span-missing-delimiter`: a code span must not wrap\n\nA backtick code span must open and close on the **same comment line**. Prose\nreflow is how this breaks: a span like\n`` `Unsupported package manager specification` `` wrapped across two comment\nlines reads perfectly balanced in source — each line's fragment looks fine —\nbut the parser sees two half-spans and emits **two**\n`tsdoc-code-span-missing-delimiter` warnings in the prod gate. When wrapping\nwould split a span, rewrap the sentence, not the span: move the whole span to\nthe next line, or shorten the prose around it. (Found live in\n`@effected\u002Fpackage-json`, where an error-message quotation\nwrapped during an ordinary TSDoc edit.)\n\nThe warning **cascades and misdirects**: once a span is unterminated, the\nbraces inside it go raw, so a single wrapped span like\n`` `PATCH \u002Frepos\u002F{owner}\u002F{repo}` `` fans out into `tsdoc-escape-right-brace`\nand `tsdoc-malformed-inline-tag` warnings whose reported lines sit BELOW the\ncomment (declaration-relative reporting). Do not chase those by escaping\nbraces — a properly closed one-line span protects `{`\u002F`}` fine, and\nbrace-bearing spans like `` `{ status: \"enabled\" }` `` are legal as-is. Find\nthe one wrapped span and rejoin it; the whole fan collapses. (Bitten twice — 14 warnings in `@effected\u002Fgit`, 5 in `@effected\u002Fgithub`, each\none wrapped argv\u002Froute literal; a reader triaging the second incident from\nthe brace warnings alone diagnosed brace-escaping, which is exactly the wrong\nfix.)\n\n## A bin-only package has no API model — `emitDts: false`, not a fake entry point\n\nA package whose published surface is a `bin` and nothing else (`exports`\nlimited to `\".\u002Fpackage.json\"`, no `index.ts`) fails the build under the\ndefault config with `Cannot merge zero API models`. Nothing is wrong\nwith the package: the declaration pass ran over zero entry points and the\nprod meta pass then had nothing to extract. The fix is **not** an `index.ts`\nthat exports nothing, and not a `meta: false` (which still runs the dts\npass). It is `emitDts = false` in the package's `effected.toml`:\n\n```toml\n# Bin-only: no declarations to bundle, no API model to extract.\n[package.build]\nemitDts = false\n```\n\n`BuildConfigInput.emitDts` is documented as the switch for *\"JS-only artifacts\nthat never consume declarations (e2e fixtures, bins, internal tools)\"* — it\nskips the dts pass (dev and prod) and therefore the meta pass, while still\nemitting JS, the byte-variant targets and the transformed `package.json`.\nSuch a package therefore carries **no `_base` suppression, no `tsdoc.json`, no\napi-extractor model and no website page** — all of that machinery reads a\n`.d.ts` that does not exist. It is also the one package shape a\ndoc-model-driven enumerator (a construct index, a docs site) must exclude by\nits `exports` map (every key `\".\u002Fpackage.json\"`), or it reports the bin as a\nperpetually missing build. The `effect-v4-cli` skill covers the package shape\nitself.\n\n## A second published entrypoint models its own surface\n\nA `package.json` `exports` map can publish more than one subpath — `.` and\n`.\u002Ftesting`, say — and api-extractor builds and gates **each entrypoint\nindependently**, as its own closed surface. A type that one entrypoint's\npublic signature reaches is not automatically visible through another\nentrypoint just because both ship from the same package: every kit type an\nentrypoint's own exported signatures name must be re-exported from **that**\nentrypoint too, or that entrypoint alone reports `ae-forgotten-export` —\n\"already exported from `.`\" is never the fix for a second entrypoint's own\ngate.\n\nWorked example: `@effected\u002Fworkspaces\u002Ftesting`\n(`packages\u002Fworkspaces\u002Fsrc\u002Ftesting.ts`). `WorkspaceLayering.checkWorkspace`'s\nrequirement and error channel reach the `WorkspaceDiscovery` service — its\nshape, options and failures. `WorkspaceDiscovery` is not part of the\npackage's main `.` surface at all; `.\u002Ftesting` re-exports it anyway, because\n`.\u002Ftesting`'s own signatures are the ones naming it. The module's own header\ncomment states the rule: \"This module is an ENTRY POINT: api-extractor\nmodels it as its own surface, so every kit type its signatures name is\nre-exported here.\"\n\n```ts\nimport type { WorkspaceDiscovery, WorkspaceLayering } from \"@effected\u002Fworkspaces\u002Ftesting\"\n\n\u002F\u002F Both resolve from the SAME subpath: WorkspaceLayering's own signatures\n\u002F\u002F reach WorkspaceDiscovery (edgesOf's parameter, checkWorkspace's\n\u002F\u002F requirement and error channel), so this entrypoint re-exports\n\u002F\u002F WorkspaceDiscovery too rather than relying on the package's main entry\n\u002F\u002F to already have done so.\ndeclare const service: WorkspaceDiscovery\ndeclare const layering: typeof WorkspaceLayering\n```\n\nA second entrypoint that re-exports types whose TSDoc `{@link}`s a class only\nthe main entry exports fails `ae-unresolved-link` on the second entry's own\nreport: the link has nothing to resolve against there. Spell those references\nin backticks instead of `{@link}`. `packages\u002Fcommands\u002FCLAUDE.md` and\n`packages\u002Fmemfs` both do this.\n\n### A subpath that names the root's types: the self-import pattern\n\nA second entrypoint whose signatures name types the root already exports —\n`@effected\u002Fnpm\u002Fnode` names `NpmConfig`, `@effected\u002Fpnpm\u002Fsettings` names\n`PnpmSettings`, `PeerDependencyRules` and `VersioningConfig` — should import\nthem from the package's own name rather than carry a copy. Three parts, all\nrequired:\n\n1. **Keep the self-import external in the declarations**, in the package's\n   `effected.toml`: `[package.build] dtsExternals = [\"@effected\u002Fnpm\"]`. The\n   subpath's `.d.ts` then refers to the root's type instead of inlining it.\n2. **Scope a suppression to that entry point only.** The second API Extractor\n   pass resolves the self-import to the built root declarations but still\n   reports the root types it reaches as forgotten, so add\n   `{ messageId = \"ae-forgotten-export\", pattern = 'entry point node\\.d\\.ts$' }`\n   (the subpath's own file name) to `[package.build.meta.tsdoc]\n   suppressWarnings`. A bare pattern would hide a real leak in the root.\n3. **Pin what the suppression could hide with a built-declaration test.** The\n   suppression is never unpinned: a declaration local to the subpath and left\n   unexported would be silenced, so a test reads the built `\u003Centry>.d.ts` and\n   asserts that every top-level declaration is exported and the import\n   specifiers are the expected ones (`__test__\u002Fnode-declarations.test.ts` in\n   `@effected\u002Fnpm`, `__test__\u002Fsettings\u002Fdeclarations.test.ts` in\n   `@effected\u002Fpnpm`). Build before the test reads it.\n\n## Reading the gate without fooling yourself\n\n`issues.json` is a **false-green oracle**. Four rules, each learned by being burned:\n\n1. **Build through Turbo.** `pnpm build --filter \u003Cpkg>` from the repo root runs\n   `types:check` first and caches. A direct `effected package build` in the\n   package directory is allowed but skips that gate and the cache, so a type\n   error the typecheck would have caught surfaces as an API Extractor failure\n   instead.\n\n2. **A crashed build can leave every bucket empty** — byte-shaped exactly like\n   a perfectly clean gate. **Read `buildOk` first**: `false` (with `failure`\n   saying why) or a missing stamp means the buckets are not evidence. An\n   unexpected `suppressed` count is a second tell: counting only\n   `source: \"api-extractor\"` entries (tsdown advisory noise such as\n   `TSGO_EXPERIMENTAL` lands there too), a package with class factories has at\n   least one `_base` entry per factory. Never conclude \"clean\"\n   from empty buckets alone; `building-a-package`'s reading-issues reference\n   has the full read order.\n\n3. **An incremental build can hide warnings.** A stale `dist\u002F.tsbuildinfo.lib`\n   feeds API Extractor an old `.d.ts`, so a warning present in a cold build is\n   absent from the incremental one. Read the gate from a cold build (`rm -rf`\n   the package's `dist` first) whenever the answer matters.\n\n4. **Concurrent agents corrupt each other's gates.** `dist\u002F` has no lock:\n   two sessions building the same package (or one running `rm -rf dist &&\n   build` while another reads) produce torn artifacts that read as rule 2's\n   clean-shaped truncation — what looks like a nondeterministic bundler flake\n   is usually contention. Treat any `issues.json` read as suspect unless it\n   immediately follows YOUR serial, cold, exit-0 build of that package. And\n   **never `pkill -f` a shared build-tool pattern** (`pkill -f \"effected package build\"`\n   killed other sessions' legitimate builds — the actual root cause of one\n   \"flake\" storm); kill your own child process by PID or not at all.\n\n`dist\u002F` is gitignored and shared. It carries no commit, no cleanliness, and no\nexit-code provenance, so it is only evidence when *your* build exited 0 against\na clean tree with nothing else building. If you mutate source to prove a fix is\nload-bearing, **rebuild afterward** — otherwise you leave an artifact that\ndescribes a tree no commit ever contained.\n\nOne cosmetic non-defect while reading built declarations: core's dollar-alias\nexport names surface verbatim — a `Schema.Array` field emits as\n`Schema.$Array\u003C…>`, `Schema.Record` as `Schema.$Record\u003C…>`. Correct output,\nobserved across multiple package builds; do not \"fix\" it, and do\nnot read it as a wrong import in the source.\n\n## Related skills\n\n`authoring-tsdoc` owns the general policy this skill assumes: the binary\nrelease-tag rule, the supported-tag allow-list, custom tags, when a\nsuppression is legitimate, the `typeof local` forgotten export, and every\nother `ae-*`\u002F`tsdoc-*` fix. `building-a-package` owns `effected.toml` and the\nbuild itself. The `tsdoc-fixer` agent carries all three and runs the fix loop\nend to end.\n\n## Rejected alternatives\n\nTwo other idioms are not used: the `@internal`-tagged base (it leaves a\nresidual non-fatal `ae-incompatible-release-tags`) and the `@public X_base`\nconst-with-explicit-annotation idiom (the inline form + scoped `_base`\nsuppression covers the recursive and `Context.Service` cases it existed for).\n\nThe inline factory + scoped `_base` suppression is the **single policy**, and\nthe transitional backlog is **drained**: no package in the kit carries a\n`@public X_base` const. `grep -rln \"_base\" packages\u002F*\u002Fsrc\u002F` returns a single\nhit — a *comment* in `runtimes\u002Fsrc\u002FNodeResolver.ts` explaining why the\nsuppression is scoped. The suppression is the build's default, so no package needs to carry the line unless it overrides\n`meta.tsdoc.suppressWarnings`. Find the packages that override it live:\n\n```bash\ngrep -l \"suppressWarnings\" packages\u002F*\u002Feffected.toml\n```\n\nEvery package that grep prints must still list the `_base` rule. If you find a\n`@public X_base` const anywhere, it is new code written against the retired\nidiom, not residue: inline the factory into the heritage clause and delete\nthe const.\n",{"name":5,"description":7},"main","0b59ef50d270bac16abbb7729cf4326d16de63f9","plugin\u002Fbuilds\u002Fcopilot\u002Fskills\u002Feffect-api-extractor-bases",2,"https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected","https:\u002F\u002Fgithub.com\u002Fspencerbeggs","2026-11-14 07:03:25",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add spencerbeggs\u002Feffected --skill effect-api-extractor-bases",{"tool":37,"command":38},"bunx","bunx skills add spencerbeggs\u002Feffected --skill effect-api-extractor-bases",{"tool":40,"command":41},"pnpm","pnpm dlx skills add spencerbeggs\u002Feffected --skill effect-api-extractor-bases","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fbuilds\u002Fcopilot\u002Fskills\u002Feffect-api-extractor-bases\u002FSKILL.md",true,[],[46,53,60,67,74,81],{"skillNo":47,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":50,"categories":51,"keywords":52,"publishedAt":19},"01M780NVTD800KMFV869YNZED6","testing-actions","Use when writing or reviewing tests for @effected\u002Fgithub-actions, @effected\u002Fgithub, @effected\u002Fsbom or @effected\u002Fcommands — stubbing the GitHub API, writing a service's layerTest or makeTest double, testing the pre\u002Fmain\u002Fpost lifecycle through Action.run, or debugging why a test hits the real network or hangs.","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Ftesting-actions\u002FSKILL.md",[],[],{"skillNo":54,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":19},"01M780NPDSGFC0FCP7696NW11J","supply-chain-attestation","Use when generating a CycloneDX SBOM, checking NTIA minimum elements, building an in-toto statement or a SLSA provenance predicate, signing into a Sigstore DSSE bundle, or uploading an attestation to GitHub.","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Fsupply-chain-attestation\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":19},"01M780NGP4QNZDH8T8Q8WZ18SA","structuring-an-action","Use when laying out a new GitHub Action repository on the @effected suite, deciding where a piece of code belongs (an entry point, a step, a shared service, a shim), or reviewing an existing action's file structure against the canonical shape.","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Fstructuring-an-action\u002FSKILL.md",[],[],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":19},"01M780NCD52H1Y03FS83WJ8BT5","running-commands-and-tools","Use when running a subprocess in Effect v4, spawning a command, capturing stdout\u002Fstderr\u002Fexit code, checking whether a CLI tool is installed or which copy to run, running a package-manager script through the project's launcher, detaching a background process, or redacting a secret from argv or captured output.","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Frunning-commands-and-tools\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":19},"01M780N823DWNRMSY5X2NT5MSR","release-and-publish","Use when publishing a package to npm, reading the npm registry from Effect v4, checking whether a version is already published, cutting a release tag or GitHub release from an action, applying a release-age gate, or deriving tracking tags for GitHub Actions distribution.","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Frelease-and-publish\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M780MXRF6H20KCESTSSFB3HQ","hardening-a-parser-port","Use when porting or writing a recursive-descent parser, lexer, or tree-walker over untrusted text in the @effected monorepo — the class of hardening the cloud reviewer scans for on every migration. Covers stack-overflow depth guards (on EVERY recursion surface, which a facade has N of — not two), numeric bound guards that must reject NaN and non-integers, code-point range checks scoped to formats with wide escapes, prototype-pollution, control-character rejection, and the invariant that malformed input must fail through the typed error channel, never as an unhandled defect.","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Fhardening-a-parser-port\u002FSKILL.md",[],[]]