feat(block): Add data enrichment block

[TheodoreSpeaks]

Summary

• Add a generic Data Enrichment workflow block (purple, generated from the enrichment registry) so workflows can run Work Email / Phone Number / Company Domain / Company Info
• New internal endpoint POST /api/tools/enrichment/run (checkInternalAuth + contract) runs the same runEnrichment provider cascade; injects workspace's hosted/BYOK key via executeTool
• Block exposes the winning provider's label on output.provider, and surfaces hosted-key cost on output.cost.total so the workflow logging session bills it
• EnrichmentRunContext.tableId/rowId made optional (workflow path has no row)
• Tool + block generated from the registry — new enrichments appear automatically

Type of Change

• New feature

Testing

Tested manually; tsc clean, biome clean, bun run check:api-validation:strict passes, bun run lint clean

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)

[vercel]

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

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
docs	Skipped		May 29, 2026 1:58am

[cursor]

PR Summary

Medium Risk
New internal tool route runs paid third-party enrichment cascades with workspace keys; auth is internal-only but misconfiguration or registry changes could affect billing and data lookups across workflows.

Overview
Adds a Data Enrichment workflow block and enrichment_run tool so workflows can run registry enrichments (work email, phone, company domain/info) using the same provider cascade as table enrichments.

The block UI, tool schema, and docs are generated from ALL_ENRICHMENTS, so new registry entries show up without hand-maintaining fields. It calls a new internal POST /api/tools/enrichment/run route (checkInternalAuth, Zod contract) that runs runEnrichment with workspaceId only—EnrichmentRunContext.tableId / rowId are now optional for this path.

runEnrichment now returns provider (winning provider label). Block/tool outputs include matched, provider, enrichment-specific fields, and cost.total on hosted-key runs for workflow billing. Shared mapFieldType maps enrichment column types to block/tool types. Docs page and API validation route count (+1) are updated.

^{Reviewed by Cursor Bugbot for commit 54fab27. Bugbot is set up for automated code reviews on this repo. Configure here.}

[greptile-apps]

Greptile Summary

This PR adds a Data Enrichment workflow block that exposes the existing enrichment registry (Work Email, Phone Number, Company Domain, Company Info) as a first-class block inside workflows, backed by a new internal endpoint POST /api/tools/enrichment/run that reuses the existing provider-cascade runner.

• New EnrichmentBlock is generated dynamically from ALL_ENRICHMENTS, so new registry entries appear automatically; per-enrichment inputs are shown/hidden via condition on the operation dropdown.
• EnrichmentRunContext.tableId/rowId are made optional so the same runner works for both the table per-row path and the workflow path (which only has workspaceId); the winning provider's label is surfaced in output.provider and hosted-key cost flows through output.cost.total for billing.

Confidence Score: 5/5

Safe to merge — the new endpoint is internal-auth gated, the contract correctly validates all inputs, and the registry-driven block generation is consistent with existing patterns.

The change is well-contained: a new internal route, a new block, and a thin tool config, all delegating to an already-tested enrichment runner. The only forward-looking concern is that the output-building loops in both the block and the tool don't guard against reserved IDs (matched, provider), which would silently corrupt output schemas if a future enrichment happened to use those names.

apps/sim/blocks/blocks/enrichment.ts and apps/sim/tools/enrichment/run.ts — both build output maps via loops that can unconditionally overwrite the statically defined matched/provider entries.

Important Files Changed

Filename	Overview
apps/sim/app/api/tools/enrichment/run/route.ts	New internal POST endpoint authenticated by checkInternalAuth; delegates to runEnrichment with the workspace's provider cascade. Logic is clean and consistent with the contract schema.
apps/sim/blocks/blocks/enrichment.ts	Block generated from the enrichment registry; unconditional loop over outputProducers can overwrite the reserved matched/provider output definitions if a future enrichment uses those IDs.
apps/sim/tools/enrichment/run.ts	Tool config for enrichment_run; …enrichmentOutputs spread last could overwrite static matched/provider output definitions if an enrichment ever uses those IDs; workspaceId is injected via _context and not a declared param.
apps/sim/enrichments/run.ts	Adds provider field to EnrichmentRunOutcome; provider.label is surfaced on hit and null on miss — clean minimal change to an existing cascade runner.
apps/sim/lib/api/contracts/tools/enrichment.ts	Zod contract for the run route; workspaceId validated with min(1), inputs defaulted to {}, response schema matches the route's return shape.

Sequence Diagram

sequenceDiagram
    participant WF as Workflow Executor
    participant Tool as enrichment_run Tool
    participant Route as POST /api/tools/enrichment/run
    participant Runner as runEnrichment()
    participant P1 as Provider 1 (e.g. Hunter)
    participant P2 as Provider 2 (e.g. PDL)

    WF->>Tool: execute(enrichmentId, inputs, _context.workspaceId)
    Tool->>Route: "POST {enrichmentId, inputs, workspaceId}"
    Route->>Route: checkInternalAuth + Zod parse
    Route->>Route: getEnrichment(enrichmentId)
    Route->>Runner: "runEnrichment(enrichment, inputs, {workspaceId})"
    Runner->>P1: executeTool(toolId, params + _context)
    alt P1 returns result
        P1-->>Runner: success + output
        Runner-->>Route: "{result, cost, error:null, provider:"Hunter"}"
    else P1 misses or errors
        P1-->>Runner: no match or error
        Runner->>P2: executeTool(toolId, params + _context)
        P2-->>Runner: success + output
        Runner-->>Route: "{result, cost, error:null, provider:"People Data Labs"}"
    end
    Route-->>Tool: "{matched, result, cost, error, provider}"
    Tool-->>WF: "{success, output:{...result, matched, provider, cost?}}"

Loading

_{Reviews (2): Last reviewed commit: "chore: re-trigger CI" | Re-trigger Greptile}

[TheodoreSpeaks]

Addressed in the latest commit:

• P1 — silent missing workspaceId: workspaceId is now required in runEnrichmentBodySchema (z.string().min(1, 'workspaceId is required')), so parseRequest returns 400 if it's omitted. Route no longer falls through to an empty string.
• P2 — docs color mismatch: BlockInfoCard color updated to #9333EA to match the block's bgColor.
• P2 — incomplete output table: docs now list matched, provider, and every per-enrichment output (email, phone, domain, industry, employeeCount, foundedYear, description).

tsc + biome + check:api-validation:strict clean.

[TheodoreSpeaks]

Addressed:

• Low — duplicated type-mapping logic: factored a shared mapFieldType helper into enrichments/providers.ts (returns the narrow 'string' | 'number' | 'boolean' | 'json' union, assignable to both OutputType and ParamType). Tool and block both import it now; the two local implementations are gone.
• Low — route baseline incremented by 2: the bot reviewed before I merged staging. After the merge, the audit is at exact match (total routes: 758 / baseline: 758), with no ratchet slack — check:api-validation:strict passes clean.

tsc + biome + audit clean.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 870203c. Configure here.}

[TheodoreSpeaks]

Addressed in latest commit — reserved output keys (matched, provider) are now assigned after the per-enrichment outputs in both the tool and the block, so a future enrichment with a colliding output id can't silently shadow them. tsc + biome clean.