feat(custom-blocks): join cross-workspace runs into the caller's trace, and map blocks per environment

[icecrasher321]

Summary

Teams orchestrating work across workspaces have two gaps that keep them on HTTP blocks instead of custom blocks: they can't see what a custom block actually did, and a forked environment silently keeps calling the environment it was forked from. This closes both.

Debugging. A custom block is an invocation boundary — a published block is org-wide, so its internals must not reach every consumer by default. The child already writes its own log row in the source workspace, correlated to the invoking run; the trace existed, it just wasn't joined. The parent's span now carries only the child's opaque execution id, and hydrateChildTraces joins the child's spans at read time, after authorizing the person reading against the child's workspace. Authorization follows the viewer rather than a flag set at publish time, re-evaluates on every read, and needs no second copy of the spans. Each hop of a nested chain is authorized against its own workspace. Boundaries left unexpanded — no access, no data, past a cap — say so, because a childless boundary span otherwise renders exactly like a leaf and a partial trace reads as a complete one.

Live runs stream too, gated on liveTraceViewerUserId, which is only set by surfaces with a single known authenticated viewer. Chat deployments stream through the same callbacks and their consumer may be anonymous, so anything that doesn't opt in keeps the boundary shut. Child spans handed to such a viewer are projected through the child's session — the invoking run's registry knows nothing about the publisher's secrets, so projecting there would leave a source-owner credential unmasked. They reach the live stream and stop: createSpanFromLog still refuses to persist them, which is what keeps read-time hydration the single authorization point.

Environments. A fork inherits its parent's organization and custom_block is keyed (organization_id, type), so a uat fork resolved to the same row and ran the prod workflow. Custom blocks become a fork-mappable resource keyed by block type — the rule every kind already follows: key by whatever the workflow references, as file does with storage keys and env-var with names. A custom block is the only resource referenced by the canvas block's own type rather than a sub-block value, so the rewrite gets its own channel. Unmapped blocks keep the source type (a type can't be emptied without deleting the node) and instead surface as unmapped and block the promote — that's what stops uat from quietly invoking prod.

Same-named environment copies now carry their source workspace, so an Access Control allowlist decision between three identical "Invoice Parser" rows is no longer a guess.

Type of Change

• New feature

Testing

bun run test (full suite), turbo run type-check (24/24), check:migrations, check:api-validation, biome — all clean on this branch rebased onto latest staging. 3482 tests across the affected surfaces.

Two full-suite failures are pre-existing and were verified as such rather than assumed:

• lib/knowledge/documents/unreadable-document.test.ts — fails against staging's own lib/knowledge/documents/ (checked out and re-run); staging's OCR rewrite broke it without updating the test.
• lib/auth/sso-trust.test.ts — a 10s-timeout test sensitive to transform-cache temperature. On a clean tree: cold cache → fails, warm cache → passes. Unrelated to this branch.

New coverage worth reviewing: hydrate-child-traces.test.ts (denied / missing / truncated / nested-per-workspace authorization, cost stripping, cycle + cap handling), the boundary tests in workflow-handler.test.ts (nothing emitted without an authorized viewer; projection goes through the child session), and trace-spans.test.ts — which fails if someone widens isWorkflowBlockType and starts persisting child spans into the parent trace.

Reviewers should focus on: the schema-quality of the read-time authorization in hydrate-child-traces.ts, the liveTraceViewerUserId gate in execution-core.ts (it is the only thing standing between a chat visitor and another workspace's block names), and the unmapped-custom-block promote blocker.

Known follow-ups (deliberately out of scope)

• Permission groups gate allowedIntegrations by block type, so a remapped fork block falls out of its groups until an admin allowlists the target type. Auto-carrying group membership across a mapping would be a privilege escalation performed on the admin's behalf, so it needs a product call. The workspace suffix added here makes the manual fix legible.
• Hydration follows 3 boundary hops vs the executor's 25-deep call chain; deeper nests say "Not expanded (nesting limit)" rather than rendering as complete.
• Parent trace detail depends on the child's log row surviving retention.

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

🤖 Generated with Claude Code

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
docs	Skipped		Aug 19, 2026 8:50pm

[cursor]

PR Summary

High Risk
Changes cross-workspace authorization for traces and live streams and fork promote semantics for custom blocks—security-sensitive boundaries where mis-gating would leak publisher block internals or leave env forks invoking prod.

Overview
Custom blocks now expose what actually ran in the caller’s trace without baking another workspace’s internals into the parent log, and forked environments can remap block types so promote doesn’t silently keep calling the source env.

Trace / debugging. Parent spans store only the child run’s opaque childExecutionId; hydrateChildTraces joins child spans at read time after checkWorkspaceAccess on each source workspace (nested hops checked separately). The trace UI explains absent children via childTraceAccess (denied, missing, truncated). Live editor runs can stream child events only when liveTraceViewerUserId is set and the viewer can read the source workspace—via emit-only liveStreamCallbacks, not parent progress markers—and child spans are projected through the child logging session. childTraceSpans are stripped from persisted traces/outputs; block logs carry childExecution for correlation.

Fork / environments. Custom blocks become a mappable fork resource keyed by block type (remapForkBlockType, promote transformBlockType). Unmapped blocks keep the source type and block promote with unmapped-custom-block messaging (distinct from “would lose”). Mapping candidates are org-scoped with workspace labels; sourceWorkspaceName disambiguates same-named blocks in Access Control and the palette.

Execution UI. isSubExecutionBlockType treats custom blocks like workflow blocks in the terminal tree; canvas drill-down still collects snapshot ids after hydration.

^{Reviewed by Cursor Bugbot for commit 9065158. Configure here.}

[greptile-apps]

Greptile Summary

The PR joins authorized cross-workspace custom-block executions into parent traces at read time and adds environment-specific custom-block mapping for workspace forks.

• Propagates opaque child execution identifiers and hydrates child traces after per-workspace authorization.
• Gates live child-span streaming on a known authenticated viewer and avoids persisting projected child spans into the parent trace.
• Adds custom-block fork resources, block-type remapping, promotion blockers, and source-workspace labels.
• Updates the database schema, migration metadata, API contracts, UI surfaces, and focused tests.

Confidence Score: 5/5

The PR appears safe to merge because no blocking failure remains within the eligible follow-up review scope.

No blocking failure remains.

Important Files Changed

Filename	Overview
apps/sim/lib/logs/execution/hydrate-child-traces.ts	Adds bounded recursive child-trace hydration with per-workspace viewer authorization, cycle handling, and explicit denied, missing, or truncated states.
apps/sim/executor/handlers/workflow/workflow-handler.ts	Propagates authorized live-viewer context across custom-block execution boundaries while projecting child spans through the child logging session.
apps/sim/lib/workflows/executor/execution-core.ts	Adds the authenticated client-session viewer identity and emit-only callbacks used to gate live cross-workspace traces.
apps/sim/ee/workspace-forking/lib/remap/remap-references.ts	Treats custom-block types as fork-remappable references and preserves unmapped source types for promotion blocking.
apps/sim/ee/workspace-forking/lib/promote/sync-blockers.ts	Adds an explicit promotion blocker for unresolved custom-block mappings.
apps/sim/ee/workspace-forking/lib/mapping/resources.ts	Adds organization-scoped custom-block mapping candidates keyed by the block type referenced by workflows.
packages/db/migrations/0293_custom_block_fork_resource.sql	Extends the workspace-fork resource enum with the custom-block resource type.
apps/sim/app/api/logs/execution/[executionId]/route.ts	Hydrates authorized child traces before collecting child workflow snapshots for execution details.

Sequence Diagram

sequenceDiagram
  participant Viewer
  participant Parent as Parent workspace
  participant Hydrator as Trace hydrator
  participant Authz as Workspace authorization
  participant Child as Child workspace
  Viewer->>Parent: Read parent execution trace
  Parent->>Hydrator: Hydrate boundary execution IDs
  Hydrator->>Child: Resolve child execution and workspace
  Hydrator->>Authz: Check viewer access to child workspace
  alt Access granted
    Authz-->>Hydrator: Granted
    Hydrator->>Child: Load and project child spans
    Child-->>Hydrator: Authorized spans
    Hydrator-->>Parent: Attach child spans
  else Access denied or data unavailable
    Authz-->>Hydrator: Denied
    Hydrator-->>Parent: Preserve opaque boundary status
  end
  Parent-->>Viewer: Authorized composite trace

Loading

_{Reviews (4): Last reviewed commit: "fix(custom-blocks): carry the emit-only ..." | Re-trigger Greptile}

[icecrasher321]

@greptile

[icecrasher321]

@cursor review

[icecrasher321]

@greptile

[icecrasher321]

@cursor review

[icecrasher321]

@greptile

[icecrasher321]

@cursor review

[cursor]

✅ Bugbot reviewed your changes and found no new issues!

Comment @cursor review or bugbot run to trigger another review on this PR

^{Reviewed by Cursor Bugbot for commit 9065158. Configure here.}