feat(core): Switch to structured destination node (no-changelog)

[mfsiega]

Summary

Internally switch to use a structured DestinationNode by making RunExecutionDataV1 the default.

Externally, nothing changes - the controller still accepts the old format and passes mode: 'inclusive' to preserve the existing behaviour. The next PR will make the change in the API and frontend to actually fix the original bug.

This PR also adds code to migrate records to the latest version of RunExecutionData, and uses this to make sure we can still read old records in the database. It adds an e2e test to verify that this happens in the execution repository.

Finally, this PR also updates the various (many) places that handle the destination node to process the new structure.

Related Linear tickets, Github issues, and Community forum posts

https://linear.app/n8n/issue/CAT-1265

Review / Merge checklist

• PR title and summary are descriptive. (conventions)
• Docs updated or follow-up ticket created.
• Tests included.
• PR Labeled with release/backport (if the PR is an urgent fix that needs to be backported)

[bundlemon]

BundleMon

Unchanged files (2)

Status	Path	Size	Limits
✅	WASM Dependencies tree-sitter-bash.wasm	181.26KB	-
✅	WASM Dependencies tree-sitter.wasm	74.47KB	-

No change in files bundle size

Groups updated (2)

Status	Path	Size	Limits
✅	**/*.js	11.38MB (+94.65KB +0.82%)	-
✅	**/*.css	231.53KB (+10.76KB +4.87%)	-

Final result: ✅

View report in BundleMon website ➡️

---

Current branch size history | Target branch size history

[codecov]

Codecov Report

❌ Patch coverage is 70.66667% with 22 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
...s/@n8n/db/src/repositories/execution.repository.ts	60.00%	9 Missing and 1 partial ⚠️
...kflow/src/run-execution-data/run-execution-data.ts	64.28%	5 Missing ⚠️
packages/cli/src/webhooks/webhook-helpers.ts	0.00%	3 Missing ⚠️
packages/cli/src/manual-execution.service.ts	33.33%	2 Missing ⚠️
packages/cli/src/webhooks/test-webhooks.ts	0.00%	2 Missing ⚠️

📢 Thoughts on this report? Let us know!

[currents-bot]

✅

E2E Tests: n8n tests passed after 9m 1.7s

🟢 588 · 🔴 0 · ⚪️ 12 · 🟣 5

View Run Details

Run Details

• Project: n8n

• Groups: 2

• Framework: Playwright

• Run Status: Passed

• Commit: 26a992d

• Spec files: 96

• Overall tests: 600

• Duration: 9m 1.7s

• Parallelization: 9

Groups

GroupId	Results	Spec Files Progress
ui	🟢 539 · 🔴 0 · ⚪️ 12 · 🟣 5	90 / 90
ui:isolated	🟢 49 · 🔴 0 · ⚪️ 0	6 / 6

---

This message was posted automatically by currents.dev | Integration Settings

[cubic-dev-ai]

cubic analysis

1 issue found across 25 files

Prompt for AI agents (all 1 issues)

Understand the root cause of the following 1 issues and fix them.


<file name="packages/cli/src/workflows/workflow-execution.service.ts">

<violation number="1" location="packages/cli/src/workflows/workflow-execution.service.ts:104">
Rule violated: **Tests**

Manual execution behavior is being rewritten to pass a structured destination node so that partial runs (per Linear issue CAT-1265) only execute the intended nodes, yet the PR adds no unit/workflow tests to demonstrate that this bug is fixed. The Community PR Guidelines explicitly require PRs to include tests, so this regression fix currently ships without the mandated coverage and risks the issue recurring.</violation>
</file>

Linked issue analysis

Linked issue: CAT-1265: Bug - Form node - executing previous nodes causes current node to run

Status	Acceptance criteria	Notes
⚠️	Executing previous nodes should not run the current node when in a 'next form step' node	Added destination.mode & conditional include, but callers still default to inclusive

_{Reply to cubic to teach it or ask questions. Re-run a review with @cubic-dev-ai review this PR}

[cubic-dev-ai]

Rule violated: Tests

Manual execution behavior is being rewritten to pass a structured destination node so that partial runs (per Linear issue CAT-1265) only execute the intended nodes, yet the PR adds no unit/workflow tests to demonstrate that this bug is fixed. The Community PR Guidelines explicitly require PRs to include tests, so this regression fix currently ships without the mandated coverage and risks the issue recurring.

Prompt for AI agents

Address the following comment on packages/cli/src/workflows/workflow-execution.service.ts at line 104:

<comment>Manual execution behavior is being rewritten to pass a structured destination node so that partial runs (per Linear issue CAT-1265) only execute the intended nodes, yet the PR adds no unit/workflow tests to demonstrate that this bug is fixed. The Community PR Guidelines explicitly require PRs to include tests, so this regression fix currently ships without the mandated coverage and risks the issue recurring.</comment>

<file context>
@@ -92,26 +93,26 @@ export class WorkflowExecutionService {
 	) {
+		const { workflowData, startNodes, dirtyNodeNames, triggerToStartFrom, agentRequest } = payload;
+		let { runData } = payload;
+		const destinationNode: IDestinationNode | undefined = payload.destinationNode
+			? {
+					nodeName: payload.destinationNode,
</file context>

[mfsiega]

This PR intentionally does not fix the e2e behaviour. A subsequent PR will update the API, as well as add tests to verify that the mode is handled appropriately.

[cubic-dev-ai]

3 issues found across 29 files

Prompt for AI agents (all 3 issues)

Understand the root cause of the following 3 issues and fix them.


<file name="packages/cli/src/events/relays/telemetry.event-relay.ts">

<violation number="1" location="packages/cli/src/events/relays/telemetry.event-relay.ts:790">
Remove the stray `console.log`—it dumps manual execution data (potentially sensitive) to stdout instead of using the structured telemetry/logging pipeline, which violates the style guidelines for production code.</violation>
</file>

<file name="packages/workflow/src/run-execution-data/run-execution-data.ts">

<violation number="1" location="packages/workflow/src/run-execution-data/run-execution-data.ts:35">
Rule violated: **Prefer Typeguards over Type casting**

`migrateRunExecutionData` now ends with a bare `as` cast, so the union `IRunExecutionDataAll` is forced into `IRunExecutionData` without a type guard or discriminant-based narrowing, violating the “Prefer Typeguards over Type casting” rule and hiding missing migrations.</violation>
</file>

<file name="packages/@n8n/db/src/repositories/execution.repository.ts">

<violation number="1" location="packages/@n8n/db/src/repositories/execution.repository.ts:221">
Rule violated: **Prefer Typeguards over Type casting**

The updated execution loaders still narrow the parsed payload with `parse(executionData.data) as IRunExecutionDataAll`, which violates the “Prefer type guards over type casting” rule. Introduce a typed variable or guard that proves the value matches `IRunExecutionDataAll` before calling `migrateRunExecutionData`, rather than asserting with `as`.</violation>
</file>

_{Reply to cubic to teach it or ask questions. Re-run a review with @cubic-dev-ai review this PR}

[cubic-dev-ai]

Rule violated: Prefer Typeguards over Type casting

migrateRunExecutionData now ends with a bare as cast, so the union IRunExecutionDataAll is forced into IRunExecutionData without a type guard or discriminant-based narrowing, violating the “Prefer Typeguards over Type casting” rule and hiding missing migrations.

Prompt for AI agents

Address the following comment on packages/workflow/src/run-execution-data/run-execution-data.ts at line 35:

<comment>`migrateRunExecutionData` now ends with a bare `as` cast, so the union `IRunExecutionDataAll` is forced into `IRunExecutionData` without a type guard or discriminant-based narrowing, violating the “Prefer Typeguards over Type casting” rule and hiding missing migrations.</comment>

<file context>
@@ -20,6 +20,17 @@ const __brand = Symbol(&#39;brand&#39;);
+		// Fall through to subsequent versions as they&#39;re added.
+	}
+	// NOTE: it&#39;s safe to assert here because we have handled all previous versions.
+	return data as IRunExecutionData;
+}
</file context>

✅ Addressed in db58367

[cubic-dev-ai]

Rule violated: Prefer Typeguards over Type casting

The updated execution loaders still narrow the parsed payload with parse(executionData.data) as IRunExecutionDataAll, which violates the “Prefer type guards over type casting” rule. Introduce a typed variable or guard that proves the value matches IRunExecutionDataAll before calling migrateRunExecutionData, rather than asserting with as.

Prompt for AI agents

Address the following comment on packages/@n8n/db/src/repositories/execution.repository.ts at line 221:

<comment>The updated execution loaders still narrow the parsed payload with `parse(executionData.data) as IRunExecutionDataAll`, which violates the “Prefer type guards over type casting” rule. Introduce a typed variable or guard that proves the value matches `IRunExecutionDataAll` before calling `migrateRunExecutionData`, rather than asserting with `as`.</comment>

<file context>
@@ -217,7 +218,7 @@ export class ExecutionRepository extends Repository&lt;ExecutionEntity&gt; {
 				return {
 					...rest,
-					data: parse(executionData.data) as IRunExecutionData,
+					data: migrateRunExecutionData(parse(executionData.data) as IRunExecutionDataAll),
 					workflowData: executionData.workflowData,
 					customData: Object.fromEntries(metadata.map((m) =&gt; [m.key, m.value])),
</file context>

[mfsiega]

This PR doesn't change the behaviour here - we were always asserting. We trust that only IRunExecutionData (of some version) has been written to the database.

[despairblue]

I made some small changes (adding tests, removing some types that can be inferred, adding a todo, etc.). Please check them out and let me know if you agree with them.

Rest look really good. Nice job!