- Pass DOMException through to the workflow VM context alongside other
Web APIs (Headers, URL, TextEncoder, etc.)
- Add DOMException reducer/reviver to the serialization pipeline,
preserving message, name, derived code, and cause when present
- Add DOMException to the Serializable type union
- 8 new tests: round-trip for AbortError, NotFoundError, default name,
cause preservation, cause absence, serialization key, cross-VM
boundaries, and VM context availability
* fix(builders): override sideEffects:false for discovered workflow/step/serde entries
When node_modules packages include "sideEffects": false in their
package.json, esbuild drops bare imports from the virtual-entry.js
file. This is incorrect because the SWC compiler transform injects
side-effectful registration code (workflow IDs, step IDs, class
serialization) into these modules.
Fix: return the resolved path alongside sideEffects: true from the
onResolve handler so esbuild uses the plugin's resolution result
instead of re-reading the package.json.
* refactor(builders): normalize sideEffectEntries with realpaths for symlink compatibility
Extract withRealpaths() helper and use it for both normalizedEntriesToBundle
and sideEffectEntries at all three bundle sites. This ensures the
sideEffects override works correctly under pnpm/workspace symlinked
layouts where enhanced-resolve may return realpaths that differ from
the original discovered file paths.
* perf(builders): skip enhanced-resolve for transitive imports when only sideEffectEntries is set
When entriesToBundle is not set (workflow/client bundles), only top-level
import statements need the sideEffects override — transitive imports
from deep within the bundle are not bare imports and don't need resolution.
Skip enhanced-resolve for non-import-statement kinds to reduce overhead.
* fix(swc-plugin): use binding name for class expression method registrations
When a pre-bundled package (e.g. via tsup) contains class expressions
like `var Foo = class _Foo { ... }`, the internal name `_Foo` is only
scoped inside the class body. The SWC plugin was incorrectly using the
internal name for method step registrations and class serialization
registrations emitted at module scope, causing ReferenceError at runtime.
Fix: always use the binding name (registration_name) for
current_class_name in visit_mut_class_expr, consistent with the existing
handling for anonymous class expressions. This ensures:
- registerStepFunction calls reference the binding name (Foo)
- Only one class registration IIFE is emitted (not duplicates for both
Foo and _Foo)
- Step IDs use the binding name in their qualified path
* refactor(swc-plugin): rename internal_class_name to tracked_class_name for clarity
The variable no longer represents the internal class expression identifier
after being reassigned to the binding name. Rename to tracked_class_name
and eliminate the intermediate registration_name variable to make the
intent clearer and reduce confusion for future readers.
* fix(swc-plugin): rewrite anonymous export default class to const declaration
When an anonymous class with serde/step methods is exported as a default
export (`export default class { ... }`), the generated registration code
(registerStepFunction, class registry IIFE) would reference a nonexistent
variable at module scope, causing a ReferenceError at runtime.
Fix: detect anonymous default class exports in visit_mut_export_default_decl
and visit_mut_export_default_expr, generate a unique binding name
(__defaultClass), and defer a rewrite in visit_mut_module_items that
transforms the export into:
const __defaultClass = class __defaultClass { ... };
export default __defaultClass;
Named default class exports (export default class Foo { ... }) are
handled by setting current_class_binding_name so the transformer
uses the existing class name for registration code.
* refactor(swc-plugin): rename __defaultClass to __DefaultClass for class naming convention
* fix(swc-plugin): fix panic for step-only anonymous default class exports
Address review feedback:
- Remove expect() that panicked when anonymous default class had step
methods but no serde methods (ident was not re-inserted). Keep
const_name in a local variable instead of relying on class_expr.ident.
- Add debug_assert for single default class export invariant.
- Update spec.md and test fixture inputs to use new this() instead of
referencing the generated binding name.
- Add step-only anonymous default class fixture to cover the bug path.
* refactor(swc-plugin): address review feedback for export default class handling
- Remove dead Expr::Class handler in visit_mut_export_default_expr
(SWC wraps parenthesized form in Expr::Paren, so it never fires)
- Extract class_needs_binding_rewrite() helper, eliminating duplicated
detection logic and unnecessary body clones
- Add debug_assert for mutual exclusivity of default_workflow_exports
and default_class_exports
- Clarify spec.md on self-name behavior difference between serde and
step-only classes
* Fix node-module-error plugin matching identifiers in multi-line comments
The findIdentifierUsage function only stripped single-line comments and
same-line block comments, but didn't track multi-line block comment state.
Lines inside JSDoc/block comments (e.g. ` * ... Writable stream`) passed
through unstripped, causing the plugin to point at comments instead of
actual code usage.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* Strip string literals before scanning for comment delimiters
Move string stripping before comment detection so that comment delimiters
inside string literals (e.g. `const s = "/*"`) don't incorrectly trigger
block comment mode. Adds regression test.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-authored-by: Peter Wielander <mittgfu@gmail.com>
* fix(builders): override sideEffects:false for discovered workflow/step/serde entries
When node_modules packages include "sideEffects": false in their
package.json, esbuild drops bare imports from the virtual-entry.js
file. This is incorrect because the SWC compiler transform injects
side-effectful registration code (workflow IDs, step IDs, class
serialization) into these modules.
Fix: return the resolved path alongside sideEffects: true from the
onResolve handler so esbuild uses the plugin's resolution result
instead of re-reading the package.json.
* refactor(builders): normalize sideEffectEntries with realpaths for symlink compatibility
Extract withRealpaths() helper and use it for both normalizedEntriesToBundle
and sideEffectEntries at all three bundle sites. This ensures the
sideEffects override works correctly under pnpm/workspace symlinked
layouts where enhanced-resolve may return realpaths that differ from
the original discovered file paths.
* perf(builders): skip enhanced-resolve for transitive imports when only sideEffectEntries is set
When entriesToBundle is not set (workflow/client bundles), only top-level
import statements need the sideEffects override — transitive imports
from deep within the bundle are not bare imports and don't need resolution.
Skip enhanced-resolve for non-import-statement kinds to reduce overhead.
* fix(swc-plugin): use binding name for class expression method registrations
When a pre-bundled package (e.g. via tsup) contains class expressions
like `var Foo = class _Foo { ... }`, the internal name `_Foo` is only
scoped inside the class body. The SWC plugin was incorrectly using the
internal name for method step registrations and class serialization
registrations emitted at module scope, causing ReferenceError at runtime.
Fix: always use the binding name (registration_name) for
current_class_name in visit_mut_class_expr, consistent with the existing
handling for anonymous class expressions. This ensures:
- registerStepFunction calls reference the binding name (Foo)
- Only one class registration IIFE is emitted (not duplicates for both
Foo and _Foo)
- Step IDs use the binding name in their qualified path
* refactor(swc-plugin): rename internal_class_name to tracked_class_name for clarity
The variable no longer represents the internal class expression identifier
after being reassigned to the binding name. Rename to tracked_class_name
and eliminate the intermediate registration_name variable to make the
intent clearer and reduce confusion for future readers.
* fix(builders): override sideEffects:false for discovered workflow/step/serde entries
When node_modules packages include "sideEffects": false in their
package.json, esbuild drops bare imports from the virtual-entry.js
file. This is incorrect because the SWC compiler transform injects
side-effectful registration code (workflow IDs, step IDs, class
serialization) into these modules.
Fix: return the resolved path alongside sideEffects: true from the
onResolve handler so esbuild uses the plugin's resolution result
instead of re-reading the package.json.
* refactor(builders): normalize sideEffectEntries with realpaths for symlink compatibility
Extract withRealpaths() helper and use it for both normalizedEntriesToBundle
and sideEffectEntries at all three bundle sites. This ensures the
sideEffects override works correctly under pnpm/workspace symlinked
layouts where enhanced-resolve may return realpaths that differ from
the original discovered file paths.
* perf(builders): skip enhanced-resolve for transitive imports when only sideEffectEntries is set
When entriesToBundle is not set (workflow/client bundles), only top-level
import statements need the sideEffects override — transitive imports
from deep within the bundle are not bare imports and don't need resolution.
Skip enhanced-resolve for non-import-statement kinds to reduce overhead.
* fix: update types and documentation for start function overloads
Ensure types are 'unknown[]' and 'unknown' for 'deploymentId' and update exports and documentation.
Slack-Thread: https://vercel.slack.com/archives/C09G3EQAL84/p1773368990070059?thread_ts=1773368990.070059&cid=C09G3EQAL84
Co-authored-by: Pranay Prakash <1797812+pranaygp@users.noreply.github.com>
* fix: use generics in deploymentId overloads to avoid contravariance issue
Addresses PR review feedback: typed workflows like
WorkflowFunction<[string], number> were not assignable to
WorkflowFunction<unknown[], unknown> under strictFunctionTypes.
Changed to generic parameters while keeping Run<unknown> return type.
Also adds type-level tests for overload resolution.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* chore: add changeset for start() deploymentId type changes
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: v0 <v0[bot]@users.noreply.github.com>
Co-authored-by: Pranay Prakash <1797812+pranaygp@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* Rename 'Workflow Development Kit' / 'DevKit' to 'Workflow SDK' across docs, code, and config
Follow-up to cdf90d5a38 (#1541)
* Fix missing </h1> closing tag and add article 'the' before 'Workflow SDK' in docs
* docs: rename 'Complex Example' to 'Instance Methods as Steps' in serialization guide
Rework the section title and introductory copy to better reflect
the purpose: making classes with Node.js APIs / side effects
workflow-compatible by adding "use step" to instance methods.
* docs: clarify that the static requirement applies to serialization hooks
Make the callout explicitly name WORKFLOW_SERIALIZE and
WORKFLOW_DESERIALIZE so it doesn't read as a blanket restriction
on instance methods, which would contradict the 'Instance Methods
as Steps' section below.
* fix: check target run capabilities before encrypting hook payloads
When resumeHook()/resumeWebhook() is called on a newer deployment that
supports encryption, it would encode the payload with the 'encr' format.
If the target workflow run was created by an older deployment that
predates encryption support, the run would fail with:
Error: Unknown serialization format: "encr". Known formats: devl
Add a capabilities table that maps @workflow/core versions to supported
serialization formats. Before encoding, resumeHook() now checks the
target run's workflowCoreVersion and suppresses encryption when the
run's deployment doesn't support it.
* address review: guard against invalid/non-string workflowCoreVersion
- Validate with semver.valid() before comparing, falling back to
baseline formats for malformed version strings
- Add typeof guard at the call site in resumeHook() since
executionContext is Record<string, any>
- Add tests for invalid version strings (dev, empty, partial, etc.)
- Add encryption commit reference to capabilities module header
In Nitro/Nuxt local production builds, the generated workflow/steps.mjs bundle could be treated as a pure export provider when the virtual handler only imports { POST }. That allowed the step bundle's top-level registerStepFunction(...) calls to disappear from the final Nitro server bundle, which led to runtime failures like Step \"...\" not found when a queued step executes.
Signed-off-by: comfuture <comfuture@gmail.com>
Co-authored-by: Peter Wielander <mittgfu@gmail.com>
* Polyfill TC39 `Uint8Array` base64/hex methods in workflow VM context
* Replace `declare global` with local type interfaces to avoid type leakage
* Document Uint8Array base64/hex methods in Workflow Globals page
Replace fragile setTimeout-based waits in sleep, step, and
events-consumer tests with withResolvers/vi.waitFor patterns,
following the approach established in #1347.
* Move SWC playground transform from server action to client-side WASM
Create a new swc-playground-wasm crate that bundles swc_ecma_parser,
swc_ecma_codegen, and the swc_workflow transform visitor into a single
WASM binary via wasm-bindgen/wasm-pack. This runs the code
transformation entirely in the browser, eliminating the server action
round-trip and serverless function cold starts on every keystroke.
- New packages/swc-playground-wasm Rust crate targeting wasm32-unknown-unknown
- Exposes transform() and transformAll() functions via wasm-bindgen
- Client loads WASM + JS glue from public/wasm/ as static assets
- Removed @swc/core and @workflow/swc-plugin server-side dependencies
- Removed serverExternalPackages/outputFileTracingIncludes Next.js config
- Added WASM loading indicator and error state in the UI
- Reduced transform debounce from 500ms to 300ms (no network latency)
* Add missing extends key to swc-playground-wasm turbo.json
* Fix build: ensure rustup is available for wasm32-unknown-unknown target
The Vercel build environment has a system Rust without rustup, so
the wasm32-unknown-unknown target can't be managed. Now the build
script checks for rustup specifically (not just cargo) and installs
it when missing, matching the pattern in swc-plugin-workflow/build.js.
* Add workflow type definitions to Monaco editor for intellisense
Auto-generate type declarations from built .d.ts files of workflow,
@workflow/core, @workflow/errors, @workflow/world, and @workflow/utils
packages. Register them with Monaco's TypeScript language service via
addExtraLib() so imports like 'workflow' resolve with full types,
eliminating red squiggles and enabling autocomplete/hover info.
* Fix Monaco type resolution: register root index.d.ts for each package
Monaco's NodeJs module resolution looks for index.d.ts at the package
root, not just in dist/. Register the main entry .d.ts content at both
paths (dist/ and root) so bare imports resolve with full type info.
Also remove the 2307 diagnostic suppression so non-existent imports
correctly show errors.
* Fix Monaco tooltip overflow by enabling fixedOverflowWidgets
* Fix hydration mismatch, Monaco type resolution paths, and WASM init warning
- Fix hydration mismatch: gate Reset button disabled state on isHydrated
so server and client render consistently during hydration
- Fix Monaco types: use bare node_modules/ paths instead of file:///
URIs for addExtraLib, which is what Monaco's NodeJs resolver expects
- Remove virtual package.json entries (Monaco doesn't use them)
- Fix wasm-bindgen init deprecation: pass object { module_or_path }
instead of bare string argument
* Fix Monaco module resolution: set model URI to file:/// and match addExtraLib paths
Monaco's TypeScript NodeJs resolver needs the editor model and the
addExtraLib entries to share the same URI scheme. Set the input editor
model path to file:///src/input.tsx and register all type declarations
at file:///node_modules/... so resolution of bare imports like
'workflow' correctly finds the virtual node_modules.
Also register a root index.d.ts for packages like @workflow/world that
lack an explicit 'types' field in their package.json exports.
* Use declare module ambient declarations for reliable Monaco type resolution
Replace the virtual node_modules filesystem approach with declare module
ambient declarations. This is the standard approach used by TypeScript
Playground and StackBlitz — it works regardless of Monaco's internal URI
scheme and module resolution quirks.
The generation script now:
- Registers all .d.ts files at file:///node_modules/<pkg>/dist/... paths
- Generates a global ambient declarations file with declare module blocks
that map bare import specifiers to their .d.ts entry points
- Includes @workflow/serde and workflow sub-exports (api, errors, observability)
- Supports configurable sub-export mappings per package
* Inline types into declare module blocks for full Monaco type support
Replace the export-from-file approach with fully inlined declare module
blocks. The script now reads each .d.ts entry point, recursively inlines
all relative imports, strips external import statements (resolved via
other declare module blocks), and produces a single ambient declarations
string.
This correctly handles:
- unique symbol exports (@workflow/serde)
- cross-package re-exports (workflow re-exporting from @workflow/core)
- JSDoc comments preserved for hover documentation
- Sub-path exports (workflow/api, workflow/errors, workflow/observability)
- Added @workflow/serde package
* Add workspace packages as dependencies so Turbo builds their types
The generate-monaco-types script reads .d.ts files from the built
dist/ directories of workflow, @workflow/core, @workflow/errors, etc.
On Vercel, Turbo only builds explicit dependencies — without these
workspace references, the packages were never built and the dist/
directories didn't exist, resulting in 0 type modules generated.
* Add @types/node declarations to Monaco editor
Register all @types/node .d.ts files via addExtraLib so Node.js
built-in modules (crypto, fs, path, etc.) are available in the
playground editor with full type information.
* Collect @types/node .d.ts files recursively to include subpath modules
The previous non-recursive scan missed subdirectory files like
fs/promises.d.ts, stream/web.d.ts, dns/promises.d.ts, etc., causing
'Cannot find module node:fs/promises' errors.
* Add AI agent detection and automatic markdown rewrites
When AI agents (Claude, ChatGPT, Cursor, etc.) request docs pages,
the proxy now detects them and transparently rewrites to the markdown
route — matching the geistdocs template default.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* Fix: Missing import for `isAIAgent` in `docs/proxy.ts` causes a ReferenceError at runtime when the AI agent detection code path is reached.
This commit fixes the issue reported at docs/proxy.ts:68
**Bug explanation:**
In `docs/proxy.ts`, the function `isAIAgent` is called on line 68 (`const agentResult = isAIAgent(request)`) within the AI agent detection block (lines 63-87). However, this function was never imported into the file. The function is defined and exported in `docs/lib/ai-agent-detection.ts`, but the import statement was omitted when the AI agent detection feature was added to `proxy.ts`.
This would cause a `ReferenceError: isAIAgent is not defined` at runtime whenever a request matches the condition on lines 64-67 (any request to `/docs` or `/docs/*` that doesn't include `/llms.mdx/`). This is a critical path — every docs page request would hit this code.
**Fix explanation:**
Added the missing import statement: `import { isAIAgent } from '@/lib/ai-agent-detection';` at line 10 of `proxy.ts`, after the existing imports. This correctly resolves the `isAIAgent` reference to the exported function in `docs/lib/ai-agent-detection.ts`.
Co-authored-by: Vercel <vercel[bot]@users.noreply.github.com>
Co-authored-by: molebox <hello@richardhaines.dev>
---------
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-authored-by: Vercel <vercel[bot]@users.noreply.github.com>
Peter Wielander