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

1 Skipped Deployment

Project	Deployment	Review	Updated (UTC)
docs	Skipped		Dec 20, 2025 3:13am

Greptile Summary

This PR enables using multiple instances of workflow executor and knowledge base tools in agent blocks. Previously, the UI prevented selecting the same tool twice, but this was overly restrictive since each instance can target a different workflow or knowledge base.

The implementation creates unique tool IDs by appending resource IDs (workflowId/knowledgeBaseId) when transforming tools for the LLM, then strips these suffixes during execution to look up the base tool configuration. This allows the LLM to distinguish between instances while the execution layer uses the shared tool definition.

Key changes:

• Modified isToolAlreadySelected to allow multiple workflow/knowledge block instances
• Added resource ID suffix to tool IDs in transformBlockTool
• Added normalizeToolId function to strip suffixes before tool lookup
• Fixed popover search state reset issue
• Updated turbo to 2.7.0

Confidence Score: 5/5

• This PR is safe to merge with minimal risk
• The implementation is well-designed with a clean separation of concerns. The tool ID transformation and normalization logic is straightforward and handles the multi-instance requirement correctly. All changes are additive and backward compatible - existing single-instance tools continue to work unchanged. The popover fix addresses a previous review comment. No security concerns or breaking changes identified.
• No files require special attention

Important Files Changed

Sequence Diagram

sequenceDiagram
    participant UI as Tool Input UI
    participant Transform as transformBlockTool
    participant Agent as Agent Handler
    participant LLM as LLM Provider
    participant Execute as executeTool
    participant Normalize as normalizeToolId
    participant Tool as Tool Config

    UI->>UI: User adds workflow/KB tool instance
    Note over UI: isToolAlreadySelected returns false<br/>for workflow & knowledge blocks
    UI->>Transform: tool with workflowId/knowledgeBaseId
    Transform->>Transform: Create unique ID<br/>(e.g., workflow_executor_uuid)
    Transform-->>Agent: {id: "workflow_executor_uuid", params, ...}
    Agent->>LLM: Send tools with unique IDs
    LLM->>LLM: Agent decides to call tool
    LLM-->>Agent: tool_call: "workflow_executor_uuid"
    Agent->>Execute: executeTool("workflow_executor_uuid", params)
    Execute->>Normalize: normalizeToolId("workflow_executor_uuid")
    Normalize-->>Execute: "workflow_executor"
    Execute->>Tool: getTool("workflow_executor")
    Tool-->>Execute: Tool config
    Execute->>Execute: Execute tool with params
    Execute-->>Agent: Tool result

@greptile

Greptile's behavior is changing!

From now on, if a review finishes with no comments, we will not post an additional "statistics" comment to confirm that our review found nothing to comment on. However, you can confirm that we reviewed your changes in the status check section.

_{This feature can be toggled off in your Code Review Settings by deselecting "Create a status check for each PR".}